Title: Modeling Moderated Networks

Version: 0.9.0

Maintainer: Trevor Swanson <trevorswanson222@gmail.com>

Description: Methods for modeling moderator variables in cross-sectional, temporal, and multi-level networks. Includes model selection techniques and a variety of plotting functions. Implements the methods described by Swanson (2020) https://www.proquest.com/openview/d151ab6b93ad47e3f0d5e59d7b6fd3d3.

License: GPL (≥ 3)

URL: https://github.com/tswanson222/modnets

BugReports: https://github.com/tswanson222/modnets/issues

Imports: abind, corpcor, ggplot2, glinternet, glmnet, gridExtra, gtools, igraph, interactionTest, leaps, lme4, lmerTest, Matrix, methods, mvtnorm, parallel, pbapply, plyr, psych, qgraph, reshape2, systemfit

Suggests: sn, arm, hierNet, psychTools

Encoding: UTF-8

LazyData: true

RoxygenNote: 7.1.1

Depends: R (≥ 2.10)

NeedsCompilation: no

Packaged: 2021年09月30日 16:28:09 UTC; t092s958@home.ku.edu

Author: Trevor Swanson [aut, cre]

Repository: CRAN

Date/Publication: 2021年10月01日 08:20:02 UTC

centAuto(x, which.net = "temporal", weighted = TRUE, signed = TRUE)
clustAuto(x, thresholdWS = 0, thresholdON = 0)

x

Output from one of the primary modnets functions. Can also supply a list of network models, and the function will be applied to all models in the list.

which.net

Only applies to SUR networks, as well as those fit with the mlGVAR function. Character string to indicate which type of network to compute centrality values for. Options are "temporal" for the temporal network, "contemporaneous" for the contemporaneous network, "PDC" for the partial directed correlation network, and "interactions" for the temporal interaction network.

weighted

Logical. If TRUE then results are converted to an unweighted network.

signed

Logical. Determines whether to ignore the signs of edges or not. Primarily affects the output for expected influence statistics.

thresholdWS

Numeric threshold for the WS values.

thresholdON

Numeric threshold for the Zhang values.

x <- fitNetwork(ggmDat, 'M')
clustAuto(x)
centAuto(x, 'interactions')

centTable(
 Wmats,
 scale = TRUE,
 which.net = "temporal",
 labels = NULL,
 relative = FALSE,
 weighted = TRUE,
 signed = TRUE
)
clustTable(Wmats, scale = TRUE, labels = NULL, relative = FALSE, signed = TRUE)

Wmats

Output from one of the primary modnets functions.

scale

Logical. Determines whether to standardize values within each measure (i.e., convert to z-scores).

which.net

Only applies to SUR networks, as well as those fit with the mlGVAR function. Character string to indicate which type of network to compute centrality values for. Options are "temporal" for the temporal network, "contemporaneous" for the contemporaneous network, "PDC" for the partial directed correlation network, and "interactions" for the temporal interaction network.

labels

Character vector to input the names of the nodes. If left NULL, the function defaults to the node names specified by the model.

relative

Logical. Determines whether to scale values within each measure relative to the largest value within that measure.

weighted

Logical. If TRUE then results are converted to an unweighted network.

signed

Logical. Determines whether to ignore the signs of edges or not. Primarily affects the output for expected influence statistics.

x <- fitNetwork(gvarDat, 'M', lags = TRUE)
clustTable(x)
centTable(x, which.net = 'interactions')

centPlot(
 Wmats,
 scale = c("z-scores", "raw", "raw0", "relative"),
 which.net = "temporal",
 include = "all",
 labels = NULL,
 orderBy = NULL,
 decreasing = FALSE,
 plot = TRUE,
 verbose = TRUE,
 weighted = TRUE,
 signed = TRUE
)
clustPlot(
 Wmats,
 scale = c("z-scores", "raw", "raw0", "relative"),
 include = "all",
 labels = NULL,
 orderBy = NULL,
 decreasing = FALSE,
 plot = TRUE,
 signed = TRUE,
 verbose = TRUE
)
plotCentrality(
 Wmats,
 which.net = "temporal",
 scale = TRUE,
 labels = NULL,
 plot = TRUE,
 centrality = "all",
 clustering = "Zhang"
)

Wmats

Output from one of the primary modnets functions.

scale

If "z-scores", then standardized values will be plotted. If "relative", then values will be scaled relative to the largest value on each measure. "raw" can be used to plot raw values.

which.net

Only applies to SUR networks, as well as those fit with the mlGVAR function. Character string to indicate which type of network to compute centrality values for. Options are "temporal" for the temporal network, "contemporaneous" for the contemporaneous network, "PDC" for the partial directed correlation network, and "interactions" for the temporal interaction network.

include

Character vector of which centrality measures to plot. "Betweenness" and "Closeness" are available for all types of network. "Strength" and "ExpectedInfluence" are only available for GGMs. And "InStrength", "OutStrength", "InExpectedInfluence", "OutExpectedInfluence" are only available for SUR networks. Defaults to "all"

labels

Character vector listing the node names. If NULL, then the names specified by the model are used.

orderBy

Character string specifying which measure to order values by.

decreasing

Logical. Only relevant if orderBy is specified. Determines whether values are organized from highest to lowest, or vice versa.

plot

Logical. Determines whether to plot the output or not.

verbose

Logical. Determines whether to return a message about the plot (messages are only shown if values are scaled).

weighted

See centTable or clustTable .

signed

See centTable or clustTable .

centrality

Character vector of centrality measures to plot. Defaults to "all".

clustering

Character vector of clustering measures to plot. Defaults to "Zhang".

x <- fitNetwork(ggmDat)
centPlot(x)
clustPlot(x)
plotCentrality(x)

modLL(
 net0,
 net1 = NULL,
 nodes = FALSE,
 lrt = NULL,
 all = FALSE,
 d = 4,
 alpha = 0.05,
 orderBy = NULL,
 decreasing = TRUE
)
SURll(
 net0,
 net1 = NULL,
 nodes = FALSE,
 lrt = NULL,
 all = FALSE,
 d = 4,
 alpha = 0.05,
 s = "res",
 orderBy = NULL,
 decreasing = TRUE,
 sysfits = FALSE
)
modTable(
 net0,
 nodes = FALSE,
 orderBy = TRUE,
 d = 4,
 alpha = 0.05,
 decreasing = TRUE,
 names = NULL,
 rmsea = FALSE
)
SURtable(
 net0,
 nodes = FALSE,
 orderBy = TRUE,
 d = 4,
 alpha = 0.05,
 decreasing = TRUE,
 names = NULL,
 rmsea = FALSE,
 s = "res"
)

net0

Output from one of the main modnets functions. Alternatively, a list of network models can be provided. This list should be named for the easiest interpretation of results.

net1

For modLL and SURll , can be used to supply a second network to compare to net0 via an LRT. Or if lrt = FALSE, then relevant statistics will be returned for both net0 and net1. Importantly, if one network is provided for net0, and another is provided for net1, then the names in the output will reflect these argument names. This can be somewhat confusing at times, so ultimately it is not recommended to use this argument. Instead, try supplying both networks (or more) as a named list to the net0 argument for the most customization.

nodes

Logical. Determines whether to compute omnibus or nodewise statistics and tests. If TRUE, then LL values for nodewise models will be returned, and any LRTs requested will reflect nodewise tests.

lrt

Logical. Determines whether to conduct an LRT or not. If FALSE, then only LL-related statistics will be returned for all models supplied. Only relevant for modLL and SURll

all

Logical. If TRUE, then omnibus LL statistics as well as nodewise statistics are returned for either LL function.

d

Number of decimal places to round outputted statistics to.

alpha

Alpha level for LRTs. Defaults to .05.

orderBy

Can be one of "LL", "df", "AIC", "BIC" to indicate which statistic to order the table by. If using modTable or SURtable , then a value of TRUE will organize the output by the LRT column, which indicates the number of times that a particular model performed better than another based on the pairwise LRTs. Higher values indicate that the model was selected more often in comparison with other models in the list.

decreasing

Logical. Determines whether to organize output from highest to lowest, or vice versa, in accordance with the value of orderBy.

s

Character string indicating which type of residual covariance matrix to compute for SUR models. Options include "res", "dfres", "sigma". "sigma" uses the residual covariance matrix as computed by the systemfit::systemfit function. "res" and "dfres" compute the matrix based directly on the residual values. "dfres" is the sample estimator that uses N - 1 in the denominator, while "res" just uses N.

sysfits

Logical, only relevant to SURll when multiple networks are included in a list. Does not currently work when there are two networks in the list, but does work with 3 or more. Returns the omnibus model statistics based on functions available to output from the systemfit::systemfit function. This allows for some additional statistics such as SSR, detSigma, OLS.R2, McElroy.R2.

names

Character vector containing the names of the models being compared. Only relevant to the modTable and SURtable . Alternatively, models can be named by supplying a named list to the net0 argument.

rmsea

Logical. Relevant to modTable and SURtable . Determines whether to return RMSEA values, as well as tests comparing RMSEA across each pair of models.

data <- na.omit(psychTools::msq[, c('hostile', 'lonely', 'nervous', 'sleepy', 'depressed')])
##### Use modLL() for GGMs
ggm1 <- fitNetwork(data[, -5])
ggm2 <- fitNetwork(data, covariates = 5)
ggm3 <- fitNetwork(data, moderators = 5)
modLL(ggm1)
modLL(ggm2)
modLL(ggm1, ggm2)
modLL(ggm1, ggm2, nodes = TRUE)
modLL(list(ggm1 = ggm1, ggm2 = ggm2))
modLL(list(GGM1 = ggm1, GGM2 = ggm2), nodes = TRUE)
ggms <- list(ggm1, ggm2, ggm3)
modLL(ggms)
modTable(ggms)
modTable(ggms, names = c("GGM1", "GGM2", "GGM3"))
names(ggms) <- c("GGM1", "GGM2", "GGM3")
modTable(ggms)
modLL(ggms)
##### Use SURll() for SUR networks
sur1 <- fitNetwork(data[, -5], lags = TRUE)
sur2 <- fitNetwork(data, covariates = 5, lags = TRUE)
sur3 <- fitNetwork(data, moderators = 5, lags = TRUE)
SURll(sur1)
SURll(sur2)
SURll(sur1, sur2)
SURll(sur1, sur2, nodes = TRUE)
SURll(list(SUR1 = sur1, SUR2 = sur2), nodes = TRUE)
surs <- list(sur1, sur2, sur3)
SURll(surs)
SURtable(surs, names = c('SUR1', "SUR2", "SUR3"))
names(surs) <- c("SUR1", "SUR2", "SUR3")
SURll(surs)
SURtable(surs)

SURfit(
 data,
 varMods = NULL,
 mod = "min",
 maxiter = 100,
 m = NULL,
 type = "g",
 center = TRUE,
 scale = FALSE,
 exogenous = TRUE,
 covs = NULL,
 sur = TRUE,
 consec = NULL,
 ...
)

data

Dataframe or matrix containing idiographic temporal data.

varMods

Output of varSelect or modSelect . The latter must be applied to resample results in order for it to work as input for this argument.

mod

Character string. Only applies if output from varSelect or modSelect is used to constrain the model, and cross-validation "CV" was set as the criterion for model/variable selection. Options include "min", which uses the lambda value that minimizes the objective function, or "1se" which uses the lambda value at 1 standard error above the value that minimizes the objective function.

maxiter

Numeric. The maximum number of iterations to attempt before stopping the function.

m

Character string or numeric value to specify the moderator (if any).

type

Indicates the type of model to use, either "g" for gaussian, or "c" for categorical (i.e., binary, at present). This argument should not be edited by the user, as the appropriate input will automatically be detected.

center

Logical. Determines whether to mean-center the variables.

scale

Logical. Determines whether to standardize the variables.

exogenous

Logical. See fitNetwork function for details.

covs

something

sur

Logical. Provides input to the method argument of the systemfit::systemfit function. If TRUE, then the method will be "SUR". If FALSE, then the method will be "OLS". These two methods only differ when constraints are applied. When a saturated model is fit, both methods produce the same results.

consec

A logical vector that identifies which values to include in accordance with the beepno and dayno arguments in the fitNetwork function.

...

Additional arguments.

SURnet(
 fit,
 dat,
 s = "sigma",
 m = NULL,
 threshold = FALSE,
 mval = NULL,
 medges = 1,
 pcor = "none"
)

fit

Output from SURfit

dat

A list containing elements "Y" and "X" elements, to reflect the outcome and predictor matrices. These are lagged data matrices, and can be automatically created through the internal modnets:::lagMat function. These transformed matrices must be supplied in conjunction with the SURfit output in order to construct network models.

s

Character string indicating which type of residual covariance matrix to compute for SUR models. Options include "res", "dfres", "sigma". "sigma" uses the residual covariance matrix as computed by the systemfit::systemfit function. "res" and "dfres" compute the matrix based directly on the residual values. "dfres" is the sample estimator that uses N - 1 in the denominator, while "res" just uses N.

m

Character string or numeric value to specify the moderator (if any).

threshold

See corresponding argument of fitNetwork

mval

Numeric. See corresponding argument of fitNetwork

medges

Numeric. See corresponding argument of fitNetwork

pcor

See corresponding argument of fitNetwork

bfiDat

bootNet(
 data,
 m = NULL,
 nboots = 10,
 lags = NULL,
 caseDrop = FALSE,
 rule = "OR",
 ci = 0.95,
 caseMin = 0.05,
 caseMax = 0.75,
 caseN = 10,
 threshold = FALSE,
 fits = NULL,
 type = "g",
 saveMods = TRUE,
 verbose = TRUE,
 fitCoefs = FALSE,
 size = NULL,
 nCores = 1,
 cluster = "mclapply",
 block = FALSE,
 maxiter = 10,
 directedDiag = FALSE,
 beepno = NULL,
 dayno = NULL,
 ...
)

data

Dataframe or matrix.

m

Numeric or character string. Indicates which variable should be treated as a moderator (if any).

nboots

Number of bootstrapped samples.

lags

Numeric or logical, to indicate whether or not a temporal network is being estimated. Maximum of 1 lag – meaningful values are either 1 or TRUE.

caseDrop

Logical. Determines whether to do a caseDrop bootstrap procedure or not.

rule

Only applies to GGMs (including between-subjects networks) when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

ci

Numeric, between 0 and 1. The level of the confidence intervals estimated. Defaults at .95

caseMin

Numeric. The minimum proportion of the sample that should be taken when caseDrop = TRUE. Provide a value between 0 and 1. The value indicates the smallest proportion of the total sample size to test in the case-dropping procedure,

caseMax

Numeric. The maximum proportion of the sample that should be taken when caseDrop = TRUE. Provide a value between 0 and 1. The value indicates the largest proportion of the total sample size to test in the case-dropping procedure,

caseN

Numeric. The number of samples to draw at each sample size tested when caseDrop = TRUE.

threshold

Logical or numeric. If TRUE, then a default value of .05 will be set. Indicates whether a threshold should be placed on the bootstrapped samples. A significant choice by the researcher. Only applies when a variable selection procedure is applied, or whether a resample object is used as input.

fits

A list of all fitted models, if available. Not likely to be used.

type

See type argument in fitNetwork function. This is where a variable selection model can be provided. This will fit the same selected model across all iterations of the bootstrapping procedure.

saveMods

Logical. Determines whether or not to return all of the fitted models – that is, all the models fit to each bootstrapped sample. Defaults to TRUE, but if FALSE then models will not be returned which can save memory.

verbose

Logical. Determines whether a progress bar should be shown, as well as whether messages should be shown.

fitCoefs

Logical, refers to the argument in the fitNetwork function. Most likely this should always be FALSE.

size

Numeric. Size of sample to use for bootstrapping. Not recommended.

nCores

If a logical or numeric value is provided, then the bootstrapping procedure will be parallelized across multiple CPUs. If numeric, this will specify the number of cores to use for the procedure. If TRUE, then the parallel::detectCores function of the parallel package will be run to maximize the number of cores available. Defaults to 1, which does not run any parallelization functions.

cluster

Character string to indicate which type of parallelization function to use, if nCores > 1. Options are "mclapply" or "SOCK".

block

Logical or numeric. If specified, then this indicates that lags != 0 or lags != NULL. If numeric, then this indicates that block bootstrapping will be used, and the value specifies the block size. If TRUE then an appropriate block size will be estimated automatically.

maxiter

The maximum number of iterations for the algorithm to go through before stopping. In some circumstances, iterated versions of the model based on subsamples of the data may not be possible to fit. In these cases, maxiter specifies the number of attempts that are made with different versions of the sample before stopping the algorithm.

directedDiag

logical

beepno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with dayno argument.

dayno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with beepno argument.

...

Additional arguments.

boot1 <- bootNet(ggmDat, 'M')
summary(boot1)
boot2 <- bootNet(gvarDat, 'M', lags = 1)
mod1 <- varSelect(gvarDat, 'M', lags = 1)
boot3 <- bootNet(gvarDat, 'M', lags = 1, type = mod1, caseDrop = TRUE)
summary(boot3)

## S3 method for class 'bootNet'
summary(object, centrality = TRUE, ...)
cscoef(object, cor = 0.7, ci = 0.95, first = TRUE, verbose = TRUE)

object

bootNet output

centrality

Logical. Determines whether or not strength centrality and expected influence should be computed for output.

...

Additional arguments.

cor

Numeric value to indicate the correlation stability value to be computed.

ci

Numeric. Confidence interval level for CS coefficient.

first

Logical. Whether or not to count the first instance that a CS statistic dips below the requisite threshold. Often times the value of this will not affect the results. When it does, if first = TRUE then the calculation will be more conservative.

verbose

Logical. Whether to write out the full statement of the CS coefficient in output. Set to FALSE if you want the details about the CS coefficient saved as attributes on the output.

boot1 <- bootNet(ggmDat, 'M')
summary(boot1)
boot2 <- bootNet(gvarDat, 'M', lags = 1)
mod1 <- varSelect(gvarDat, 'M', lags = 1)
boot3 <- bootNet(gvarDat, 'M', lags = 1, type = mod1, caseDrop = TRUE)
summary(boot3)

compareVAR(m1, m2, m3 = NULL, anova = NULL, type = "tempMods")

m1

Output from lmerVAR .

m2

Output from another run of lmerVAR . Necessary to supp

m3

Output from a third run of lmerVAR . This is optional.

anova

If NULL, then the results of each nodewise comparison will be displayed. If numeric, then this indicates which nodewise comparison to home in on. anova = 1 will show the full ANOVA results for the first predictor. anova = 2 will show the full ANOVA results for the second predictor, etc.

type

Character string. Either "tempMods" or "contempMods". Determines whether to compare the temporal network outputs or the contemporaneous network outputs with ANOVA.

fit1 <- lmerVAR(mlgvarDat, temporal = "fixed", contemp = "orthogonal")
fit2 <- lmerVAR(mlgvarDat, temporal = "orthogonal", contemp = "orthogonal")
compareVAR(fit1, fit2)

condPlot(
 out,
 to,
 from,
 swap = FALSE,
 avg = FALSE,
 compare = NULL,
 hist = FALSE,
 xlab = NULL,
 mods = NULL,
 nsims = 500,
 xn = NULL,
 getCIs = FALSE,
 discrete = FALSE,
 ylab = NULL,
 main = NULL,
 midline = TRUE
)

out

Output from fitNetwork or resample . Can also provide the fixedNets or betweenNet element of the mlGVAR output.

to

Outcome variable, specified with character string or numeric value.

from

Predictor variable, specified with character string or numeric value.

swap

Logical. Serves to switch the arguments for to and from.

avg

Logical. If TRUE then the average relationship between the two variables is displayed. Only works for GGMs.

compare

Two values can be supplied to indicate levels of the moderator to be compared.

hist

Logical. Determines whether to show a histogram of the data distribution at the bottom of the plot.

xlab

Character string for labeling the x-axis.

mods

This argument will be removed. Model output is automatically detected based on fit argument.

nsims

Number of iterations to simulate the posterior distribution.

xn

Numeric value to indicate how many values of the moderator should be evaluated.

getCIs

Logical. Only applies when avg = TRUE. If getCIs = TRUE, then the confidence intervals for the average difference between the maximum and minimum of the moderator will be returned.

discrete

Logical. Determines whether to treat the moderator as a discrete or continuous variable.

ylab

Character string for labeling the y-axis.

main

Character string for labeling the title of the plot.

midline

Logical. Only applies when discrete = TRUE. Shows a line at the average level of the outcome.

fit <- fitNetwork(ggmDat, 'M')
condPlot(fit, to = 'V5', from = 'V4')
condPlot(fit, to = 2, from = 3, avg = TRUE)

fitNetwork(
 data,
 moderators = NULL,
 type = "gaussian",
 lags = NULL,
 seed = NULL,
 folds = 10,
 gamma = 0.5,
 which.lam = "min",
 rule = "OR",
 threshold = FALSE,
 scale = FALSE,
 std = TRUE,
 center = TRUE,
 covariates = NULL,
 verbose = FALSE,
 exogenous = TRUE,
 mval = NULL,
 residMat = "sigma",
 medges = 1,
 pcor = FALSE,
 maxiter = 100,
 getLL = TRUE,
 saveMods = TRUE,
 binarize = FALSE,
 fitCoefs = FALSE,
 detrend = FALSE,
 beepno = NULL,
 dayno = NULL,
 ...
)

data

n x k dataframe or matrix.

moderators

Numeric or character vector indicating which variables (if any) to use as moderators.

type

Primarily used to supply a variable selection object, such as those created with varSelect or modSelect , or to indicate that a variable selection method should be employed by setting the value to "varSelect". Currently doesn't support setting the value to "resample", although this will be implemented in the future. Alternatively, this can be used to specify the type of variable for each node. In this case it should be either a single value – "gaussian" or "binomial" – or can be a vector of length k to specify which of those two types apply to each variable. These dictate which family to use for the call to stats::glm . Cannot use binomial models for SUR networks.

lags

Logical or numeric, to indicate whether to fit a SUR model or not. Set to TRUE or 1 for a SUR model fit to temporal data for a single subject.

seed

Only useful if type = "varSelect", and if the varSeed argument is not specified in the ...

folds

Can be used to specify the number of folds in cross-validation when type = "varSelect" and criterion = "CV". Overwritten if nfolds argument is provided.

gamma

Only useful if type = "varSelect" and the criterion is set to "EBIC". This is the hyperparameter for the calculation of EBIC.

which.lam

Only useful if criterion = "CV", or if a variable selection object based on cross-validation is supplied for type. Options include "min", which uses the lambda value that minimizes the objective function, or "1se" which uses the lambda value at 1 standard error above the value that minimizes the objective function.

rule

Only applies to GGMs (including between-subjects networks) when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

threshold

Determines whether to employ a p-value threshold on the model. If TRUE then this defaults to .05. Not recommended, as thresholds can be applied post-hoc through the plotting functions, or via the net and netInts functions. Recommended to leave as FALSE.

scale

Determines whether to standardize all variables or not.

std

Only applies to SUR networks. Logical. Provides input to the method argument of the systemfit::systemfit function. If TRUE, then the method will be "SUR". If FALSE, then the method will be "OLS". These two methods only differ when constraints are applied. When a saturated model is fit, both methods produce the same results.

center

Determines whether to mean-center variables or not.

covariates

Either a numeric value or character string – this could also be a vector – to indicate which variables (if any) should be treated as covariates in the model.

verbose

Logical. Determines whether to return information about the progress of the model fitting – especially when variable selection is employed – as well as prints the amount of time it takes to fit the model to the console.

exogenous

Logical. Indicates whether moderator variables should be treated as exogenous or not. If they are exogenous, they will not be modeled as outcomes/nodes in the network. If the number of moderators reaches k - 1 or k, then exogenous will automatically be FALSE.

mval

Numeric value to set the moderator variable to when computing model coefficients. Useful to create conditional networks – i.e., those whose values are conditioned on specific values of the moderator. Excellent when the moderator is a categorical variable, or when it's desired to have model estimates at +/- 1 SD around the mean of the moderator. These values must be supplied explicitly. Can only specify a single value for a given model.

residMat

Character string indicating which type of residual covariance matrix to compute for SUR models. Options include "res", "dfres", "sigma". "sigma" uses the residual covariance matrix as computed by the systemfits package. "res" and "dfres" compute the matrix based directly on the residual values. "dfres" is the sample estimator that uses N - 1 in the denominator, while "res" just uses N. Input for SURnet function.

medges

Only relevant when lags = 1 and exogenous = FALSE. Determines the linetype of moderated edges (corresponds to the lty argument of plot()).

pcor

Logical. Determines whether to operationalize the adjacency matrix as the partial correlation matrix of the data, or to use nodewise estimation. Only relevant for unmoderated networks.

maxiter

See argument of SURfit() function.

getLL

Logical. Determines whether to return log-likelihood statistics with model results. Recommended to keep TRUE.

saveMods

Logical. Determines whether to save the fitobj element of the output, which contains the nodewise models, or the SUR model output of systemfit::systemfit .

binarize

Logical. Determines whether to convert the output to a binary, unweighted network. Only relevant for GGMs.

fitCoefs

Determines whether to use the getFitCIs function on the output. Not recommended to use. The downside is that this will overwrite the fitobj element of the output which contains the actual models. Better to leave this as FALSE, and then use the getFitCIs function on the object separately.

detrend

Logical. Determines whether to remove linear trends from time series variables. Only applies to temporal networks.

beepno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with dayno argument. Only relevant to temporal data.

dayno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with beepno argument. Only relevant to temporal data.

...

Additional arguments.

fit1 <- fitNetwork(ggmDat)
fit2 <- fitNetwork(ggmDat, 'M', type = 'varSelect', criterion = 'BIC')
fit3 <- fitNetwork(gvarDat, 'M', lags = 1)

getFitCIs(fit, allNames = NULL, alpha = 0.05)

fit

Output from fitNetwork , or either the fixedNets or betweenNet element of the output from mlGVAR

allNames

Character vector containing all the predictor names. Do not change, as these are automatically detected.

alpha

Type 1 error rate. The complement of the confidence level.

x <- fitNetwork(ggmDat)
getFitCIs(x)

ggmDat

gvarDat

intsPlot(out, y = "all", nsims = 500, alpha = 0.05)

out

GGM moderated network output from fitNetwork , or output from a moderated between-subjects network fit with mlGVAR (e.g., when bm = TRUE).

y

Character string. The name of the outcome variable for which to create the plot. If y = "all", then all interaction terms associated with all outcomes will be plotted.

nsims

The number of simulations to estimate the posterior distribution of the difference between high and low levels of the confidence interval.

alpha

Alpha level that is used to compute confidence intervals.

fit <- fitNetwork(ggmDat, 'M')
plot(fit, 'ints', y = 'all')

lmerVAR(
 data,
 m = NULL,
 temporal = "default",
 contemp = "default",
 idvar = "ID",
 intvars = NULL,
 center = TRUE,
 scale = TRUE,
 centerWithin = TRUE,
 scaleWithin = FALSE,
 exogenous = TRUE,
 covariates = NULL,
 fix = NULL,
 verbose = TRUE,
 beepno = NULL,
 dayno = NULL,
 deleteMissing = TRUE
)

data

n x k dataframe or matrix.

m

Character vector or numeric vector indicating the moderator(s), if any. Can also specify "all" to make every variable serve as a moderator, or 0 to indicate that there are no moderators. If the length of m is k - 1 or longer, then it will not be possible to have the moderators as exogenous variables. Thus, exogenous will automatically become FALSE.

temporal

Only affects the model for the temporal network and between-subjects network (which is derived from the temporal network). Options are "default", "correlated", "orthogonal", "fixed", "intfixed". "correlated" makes it so that all random-effect terms are correlated, and "orthogonal" makes it so they are not. "fixed" makes it so that there is only a random intercept, but no other random-effect terms related to the individual predictors. "intfixed" essentially mimics "orthogonal", with the exception that no interaction terms have random slopes. "default" will automatically set the value to "correlated" if there are 6 or fewer nodes in the network, and "orthogonal" otherwise. The reason for this is that models with correlated random effects take substantially longer to fit than those with orthogonal effects. The "default" option is designed to strike a balance between comprehensiveness and efficiency for the average user. It is recommended to set this value manually in order to produce results according to one's individual specifications.

contemp

Options are "default", "correlated", "orthogonal". "correlated" makes it so that random-effect terms are correlated, and "orthogonal" makes it so they are not. "default" will automatically set the value to "correlated" if there are 6 or fewer nodes in the network, and "orthogonal" otherwise. The reason for this is that models with correlated random effects take substantially longer to fit than those with orthogonal effects. The "default" option is designed to strike a balance between comprehensiveness and efficiency for the average user. It is recommended to set this value manually in order to produce results according to one's individual specifications.

idvar

Character string to indicate which variable contains the participant identification numbers.

intvars

Character vector to indicate which interaction terms to include in the model. Not necessary, but useful to add significant customization and explicitly state which interactions to include in the model.

center

Logical. Determines whether to mean-center the variables.

scale

Logical. Determines whether to standardize the variables.

centerWithin

Following the application of center and scale, this determines whether to center variables within individual subjects to create subject-centered values.

scaleWithin

Following the application of center and scale, this determines whether to scale variables within individual subjects to create subject-standardized values.

exogenous

Logical. Indicates whether moderator variables should be treated as exogenous or not. If they are exogenous, they will not be modeled as outcomes/nodes in the network. If the number of moderators reaches k - 1 or k, then exogenous will automatically be FALSE.

covariates

See corresponding argument in fitNetwork function. Can supply a numeric value or vector to indicate which variables are covariates, or can supply a list containing the individual covariates separately from the dataset.

fix

Character vector to indicate which variables to only create fixed effects terms for.

verbose

Logical. Determines whether to output progress bars and messages in the console during the fitting process.

beepno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with dayno argument.

dayno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with beepno argument.

deleteMissing

Logical. Determines whether to automatically perform listwise deletion if there are any missing values in the dataset.

# The options were chosen so that the function would take less time to run
x <- lmerVAR(mlgvarDat, 'M', temporal = "fixed", contemp = "orthogonal")

mlGVAR(
 data,
 m = NULL,
 selectFUN = NULL,
 subjectNets = FALSE,
 idvar = "ID",
 exogenous = TRUE,
 center = TRUE,
 scale = TRUE,
 fixedType = "g",
 betweenType = "g",
 centerWithin = TRUE,
 scaleWithin = FALSE,
 rule = "OR",
 threshold = "none",
 verbose = TRUE,
 pcor = FALSE,
 fixedArgs = NULL,
 betweenArgs = NULL,
 bm = FALSE,
 beepno = NULL,
 dayno = NULL,
 deleteMissing = TRUE,
 ...
)

data

n x k dataframe or matrix

m

Character vector or numeric vector indicating the moderator(s), if any. Can also specify "all" to make every variable serve as a moderator, or 0 to indicate that there are no moderators. If the length of m is k - 1 or longer, then it will not be possible to have the moderators as exogenous variables. Thus, exogenous will automatically become FALSE.

selectFUN

Choose a variable selection function. Can specify either "varSelect" or "resample" to use the corresponding functions. If you want to use the resample function though, then it is recommended to specify selectFUN as one of: "stability", "split", "bootstrap" in order to identify the specific method. If selectFUN = "resample", then it is recommended to add the sampMethod argument to the call to mlGVAR .

subjectNets

If TRUE, then subject-specific networks are fit for all subjects and returned in the final output. Otherwise, can specify a single value or a vector of values to represent which subjects to return individual networks for – specifically, the SUR network. One caveat is that variable selection methods are not applied to these subject-specific networks. Further modeling could be done using the output, however.

idvar

Character string to indicate which variable contains the participant identification numbers.

exogenous

Logical. Indicates whether moderator variables should be treated as exogenous or not. If they are exogenous, they will not be modeled as outcomes/nodes in the network. If the number of moderators reaches k - 1 or k, then exogenous will automatically be FALSE.

center

Logical. Determines whether to mean-center the variables.

scale

Logical. Determines whether to standardize the variables.

fixedType

If logical, then any variable selection procedure specified by selectFUN will not be applied to the SUR network. Alternatively, a variable selection result, such as the output from either varSelect or modSelect , can be supplied to choose a specific constrained model in advance.

betweenType

If logical, then any variable selection procedure specified by selectFUN will not be applied to the SUR network. Alternatively, a variable selection result, such as the output from either varSelect or modSelect , can be supplied to choose a specific constrained model in advance.

centerWithin

Following the application of center and scale, this determines whether to center variables within individual subjects to create subject-centered values.

scaleWithin

Following the application of center and scale, this determines whether to scale variables within individual subjects to create subject-standardized values.

rule

Only applies to the between-subject network when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

threshold

Logical or numeric. If TRUE, then a default value of .05 will be set. Indicates whether a threshold should be placed on the models at each iteration of the sampling. A significant choice by the researcher.

verbose

Logical. Determines whether to output progress bars and messages in the console during the fitting process.

pcor

See corresponding argument in the fitNetwork function

fixedArgs

A named list of arguments for the variable selection function can be provided here, specifically those that are meant to be applied to the SUR network estimation.

betweenArgs

A named list of arguments for the variable selection function can be provided for the between-subjects network.

bm

Logical. Determines whether the same moderators are applied in the between-subjects network. By default, the value of m only applies to the SUR network. This allows one to decide whether or not to apply those moderators in the between-subject network.

beepno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with dayno argument.

dayno

Character string or numeric value to indiciate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with beepno argument.

deleteMissing

Logical. Determines whether to automatically perform listwise deletion if there are any missing values in the dataset.

...

Additional arguments.

fit1 <- mlGVAR(mlgvarDat, 'M')
fit2 <- mlGVAR(mlgvarDat, 'M', bm = TRUE) # Fit the same moderator in the between-subjects network
fit3 <- mlGVAR(mlgvarDat, 'M', selectFUN = 'varSelect')

mlGVARsim(
 nTime = 50,
 nPerson = 10,
 nNode = 3,
 m = NULL,
 m2 = 0.25,
 m1 = 0.7,
 m0 = 1,
 lag = 1,
 thetaVar = NULL,
 mu_SD = NULL,
 init_beta_SD = NULL,
 fixedMuSD = 1,
 shrink_fixed = 0.9,
 propPos = 0.5,
 m1SD = 0.1,
 m2SD = 0.1,
 m1_range = NULL,
 m2_range = NULL,
 shrink_deviation = 0.9,
 getM = FALSE,
 contemporaneous = "wishart",
 GGMsparsity = 0.5,
 mcenter = TRUE,
 skew = FALSE,
 skewErr = FALSE,
 ordinal = FALSE,
 nLevels = 5,
 ordWithin = TRUE,
 minOrd = 3,
 thresholds = NULL,
 mseed = NULL,
 onlyNets = FALSE,
 modType = "none"
)

nTime

Numeric value. The number of timepoints to simulate for each individual.

nPerson

The number of subjects to create data for. Can set to 1 to just simulate a single graphical VAR network.

nNode

The number of nodes/variables to simulate. Does not include a moderator if one is specified.

m

Logical. If TRUE, then a moderator variable will be simulated. Various options also available for highly specific moderator specification: "fixed", "random", "mixed1", "mixed2", "ar", "binary", "skewed", "random0", "ordinal".

m2

Numeric. If m2 >= 1, then this will determine the number of interaction effects between the moderator and some node in the network. If a value between 0 and 1 is provided, then this determines the probability of any given edge being moderated by the moderator.

m1

Functions similarly to m2, except that this argument refers to the number/probability of main effects of the moderator on any given node.

m0

Only relevant when m = "ar". Determines the autoregressive coefficient in the estimated models. Defaults to .3

lag

Numeric value, supposed to indicate the number of lags to simulate models parameters for. Recommended to leave at 1.

thetaVar

Numeric vector containing the variance associated with each node (excluding the moderator) in the contemporaneous network. If NULL, then it is assumed that the variance for each term is 1.

mu_SD

Numeric vector of length 2. The first value determines the standard deviation of the means associated with the temporal data, and the second value determines the standard deviations of the means associated with the between-subjects network.

init_beta_SD

Similar to mu_SD except that it applies to the coefficient estimates.

fixedMuSD

Standard deviation of the random values for the means of the fixed effects.

shrink_fixed

Numeric value to determine the factor by which to shrink sampled beta coefficients for fixed effects. Value between 0 and 1, where higher values are recommended.

propPos

The proportion of edges with a positive sign.

m1SD

Standard deviation of the moderator main effect coefficients.

m2SD

Standard deviation of the moderator interaction effect coefficients.

m1_range

Numeric vector of length 2. The range of values for moderator main effect coefficients.

m2_range

Numeric vector of length 2. The range of values for moderator interaction effect coefficients.

shrink_deviation

Numeric value to determine the factor by which to shrink contemporaneous coefficients. Value between 0 and 1, where higher values are recommended.

getM

If TRUE, only the data for the moderator, the moderator main effects, and interaction effects are returned.

contemporaneous

Options include "wishart", "randomGGM", "fixed". Determines how the contemporaneous network is sampled. The former two options sample different matrices for each subject, whereas "fixed" only samples one matrix and uses it for all subject contemporaneous networks.

GGMsparsity

Numeric value between 0 and 1. Determines the sparsity of sampled network matrices.

mcenter

If TRUE then the moderator variable is mean-centered.

skew

If TRUE then random values will be generated to represent the skewness of the node distributions. Alternatively, a numeric vector of length nNode can be provided to specify the skewness of each variable.

skewErr

The skewness parameter for the alpha argument in the sn::rmsn function.

ordinal

Logical. Determines whether to sample ordinal variables. If a numeric value is provided, then this will automatically be assigned to the nLevels argument.

nLevels

Number of levels for the ordinal variables. Only relevant if ordinal is not FALSE.

ordWithin

If TRUE, then variables will ordinalized within subjects, rather than across subjects.

minOrd

The minimum number of unique values allowed for each variable.

thresholds

List of length k, where each element is a numeric vector of length (nLevels - 1) containing the splitpoints for grouping each variable into ordered categories.

mseed

Numeric value for the seed to be set when

onlyNets

If TRUE then only the network models are returned, without the data. Could be used to create random models and then simulate data by another method.

modType

Determines the type of moderation to employ, such as "none", "full", "partial". See simNet

set.seed(1)
x <- mlGVARsim(nTime = 50, nPerson = 10, nNode = 3, m = TRUE)

mlgvarDat

mnetPowerSim(
 niter = 10,
 N = 100,
 p = 5,
 m = FALSE,
 m1 = 0,
 m2 = 0.1,
 sparsity = 0.5,
 lags = NULL,
 trueNet = NULL,
 threshold = TRUE,
 rule = "OR",
 avg = TRUE,
 maxiter = 100,
 saveFits = TRUE,
 saveData = FALSE,
 intercepts = NULL,
 mbinary = FALSE,
 select = NULL,
 vargs = list(),
 type = "g",
 gibbs = TRUE,
 ordinal = FALSE,
 mord = FALSE,
 nLevels = 5,
 minOrd = 3,
 div = 1000,
 modType = "none",
 m1_range = NULL,
 m2_range = c(0.1, 0.3),
 time = TRUE,
 skewErr = FALSE,
 nCores = 1,
 cluster = "mclapply",
 fixedPar = NULL,
 V2 = 1,
 ...
)

niter

Number of iterations/samples to take for each combination of parameters.

N

Numeric value, or vector of sample sizes to generate data with.

p

Numeric value, or vector of network sizes.

m

If a value is provided then a moderated network will be simulated. See simNet for details.

m1

Functions similarly to m2, except that this argument refers to the number/probability of main effects of the moderator on any given node.

m2

Numeric. If m2 >= 1, then this will determine the number of interaction effects between the moderator and some node in the network. If a value between 0 and 1 is provided, then this determines the probability of any given edge being moderated by the moderator.

sparsity

Numeric value between 0 and 1. Determines the sparsity of sampled network matrices.

lags

Determines whether the network should be a temporal network or not. If simulating a temporal network, set to TRUE or 1.

trueNet

The adjacency matrix of the data-generating network model, or a list containing the adjacency matrix as the first element, and the interaction matrix as the second element.

threshold

See corresponding argument in fitNetwork . Automatically set to TRUE if select is not NULL.

rule

Only applies to GGMs (including between-subjects networks) when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

avg

See corresponding argument of netInts

maxiter

If a model fails to be fit, this determines the maximum number of iterations to re-try it before giving up. Will also simulate new datasets at each iteration.

saveFits

Logical. Determines whether to save the models fit to each dataset at each iteration.

saveData

Logical. Determines whether to save the datasets generated at each iteration.

intercepts

A vector of means for sampling node values.

mbinary

Logical. Determines whether the moderator should be a binary variable.

select

Identifies a variable selection function – either varSelect or resample – to use for introducing variable selection at each iteration. The usefulness of this is to mimic a real-world situation, wherein the researcher may be interested in seeing how well datasets of different sizes afford models that approximate a true model after employing iterated variable selection. If TRUE then this defaults to "varSelect". Highly recommended to use the vargs argument to supply necessary information about the parameters of the variable selection process, such as sampMethod, criterion, etc.

vargs

A named list of arguments relevant to the variable selection procedure specified by the select argument.

type

Can supply a variable selection object, such as the output from either varSelect or modSelect , can be supplied to choose a specific constrained model to fit on all iterations. This is essentially an alternative to select, in that select performs variable selection at each iteration, whereas this argument defines a constrained model that is applied at every iteration.

gibbs

If TRUE, then Gibbs sampling will be used. Otherwise, data are generated from the mvtnorm::rmvnorm function based on the partial correlation matrix that is created.

ordinal

Logical. Determines whether to generate ordinal values or not.

mord

Logical. Determines whether the moderator variable should be simulated as ordinal.

nLevels

Number of levels for the ordinal variables. Only relevant if ordinal is not FALSE.

minOrd

The minimum number of unique values allowed for each variable.

div

A value to use as a sign that the sampler diverged. Can be increased based on expected range of values. If a datapoint is larger than div, then the sampler will stop.

modType

Determines the type of moderation to employ, such as "none", "full", "partial". See simNet for details.

m1_range

Numeric vector of length 2. The range of values for moderator main effect coefficients.

m2_range

Numeric vector of length 2. The range of values for moderator interaction effect coefficients.

time

If TRUE then the time it takes to simulate the data is printed to screen at the end of the sampling.

skewErr

The skewness parameter for the alpha argument in the sn::rmsn function. Only relevant when gibbs = FALSE and no moderator is specified.

nCores

Numeric value indicating the number of CPU cores to use for the resampling. If TRUE, then the parallel::detectCores function will be used to maximize the number of cores available.

cluster

Character vector indicating which type of parallelization to use, if nCores > 1. Options include "mclapply" and "SOCK".

fixedPar

Numeric. If provided, then this will be set as the coefficient value for all edges in the network. Provides a way to standardize the parameter values while varying the sparsity of the network. If length(fixedPar) == 1, then the same value will be used for all parameters. If length(fixedPar) == 2, then the first value will be for pairwise relationships, and the second value will be for interaction terms.

V2

If V2 = 1 and m2 is between 0 and 1, the number of interaction terms in the model will be determined by multiplying m2 with the number of elements in the interaction matrix and taking the ceiling.

...

Additional arguments.

x <- mnetPowerSim(niter = 10, N = c(100, 200))
summary(x)
plot(x)

modSelect(
 obj,
 data = NULL,
 fit = FALSE,
 select = "select",
 thresh = NULL,
 ascall = TRUE,
 type = "gaussian",
 ...
)

obj

resample output

data

The dataframe used to create the resample object. Necessary if ascall = TRUE or fit = TRUE.

fit

Logical. Determines whether to fit the selected model to the data or just return the model specifications. Must supply a dataset in the data argument as well.

select

Character string, referring to which variable of the output should be used as the basis for selecting variables. If the resampling method was either "bootstrap" or "split", then setting select = "select" will select variables based on the aggregated p-values being below a pre-specified threshold. Setting select = "select_ci", however, will use the adjusted confidence intervals rather than p-values to select variables. Alternatively, if select = "freq" then the thresh argument can be used to indicate the minimum selection frequency across iterations. In this case, variables are selected based on how frequently they were selected in the resampling procedure. This also works if select is simply set a numeric value (this value will serve as the value for thresh).

When the resampling method was "stability", the default option of select = "select" chooses variables based on the original threshold provided to the resample function, and relies on the simultaneous selection proportion (the "freq" column in the "results" element). Alternatively, if select is a numeric value, or a value for thresh is provided, that new frequency selection threshold will determine the choice of variables. Alternatively, one can specify select = "split1" or select = "split2" to base the threshold on the selection frequency in one of the two splits rather than on the simultaneous selection frequency which is likely to be the most conservative.

For all types of resample objects, when select = "Pvalue" then thresh can be set to a numeric value in order to select variables based on aggregated p-values. For the "bootstrapping" and "split" methods this allows one to override the original threshold (set as part of resample ) if desired.

thresh

Numeric value. If select = "Pvalue", then this value will be the p-value threshold. Otherwise, this value will determine the minimum frequency selection threshold.

ascall

Logical. Determines whether to return a list with arguments necessary for fitting the model with do.call to fitNetwork . Only possible if a dataset is supplied.

type

Should just leave as-is. Automatically taken from the resample object.

...

Additional arguments.

res1 <- resample(ggmDat, m = 'M', niter = 10)
mods1 <- modSelect(res1)
fit1 <- fitNetwork(ggmDat, morderators = 'M', type = mods1)
res2 <- resample(ggmDat, m = 'M', sampMethod = 'stability')
fit2 <- modSelect(res2, data = ggmDat, fit = TRUE, thresh = .7)

Model fitting: fitNetwork, mlGVAR, lmerVAR

Variable selection: varSelect, resample, modSelect

Model comparison: modLL, modTable, SURll, SURtable

Simulating network data: simNet, mlGVARsim, mnetPowerSim

Centrality and clustering: centTable, clustTable

Adjacency matrices: net, netInts

Bootstrapping and stability: bootNet

Package: modnets

Title: Modeling Moderated Networks

Version: 0.9.0

Author: Trevor Swanson

Maintainer: trevorswanson222@gmail.com

URL: https://github.com/tswanson222/modnets

BugReports: https://github.com/tswanson222/modnets/issues

License: GPL-3

Imports: abind, corpcor, ggplot2, glinternet, glmnet, gridExtra, gtools, igraph, interactionTest, leaps, lme4, lmerTest, Matrix, methods, mvtnorm, parallel, pbapply, plyr, psych, qgraph, reshape2, systemfit

Suggests: sn, arm, hierNet, psychTools

Date: 2021年09月26日

net(
 fit,
 n = "beta",
 threshold = FALSE,
 rule = "OR",
 binary = FALSE,
 nodewise = FALSE,
 d = 14,
 r = NULL
)
netInts(
 fit,
 n = "temporal",
 threshold = FALSE,
 avg = FALSE,
 rule = "none",
 r = NULL,
 empty = TRUE,
 mselect = NULL
)

fit

A fitted network model. Can be the output from fitNetwork , mlGVAR , lmerVAR , bootNet , resample , simNet , or mlGVARsim .

n

When multiple networks exist for a single object, this allows the user to indicate which adjacency matrix to return. For a GGM, all values of this argument return the same adjacency matrix. For a SUR network, "beta" and "temporal" return the coefficients associated with the temporal network, while "pdc" returns the Partial Directed Correlations, or the standardized temporal network. "contemporaneous" and "pcc" return the standardized contemporaneous network (Partial Contemporaneous Correlations). "kappa" returns the unstandardized residual covariance matrix. All of these terms apply for multilevel networks, but "between" can also return the between-subjects network. If a numeric or logical value is supplied, however, this argument will function as the threshold argument. A numeric value will set a threshold at the supplied value, while TRUE will set a threshold of .05.

threshold

A numeric or logical value to set a p-value threshold. TRUE will automatically set the threshold at .05.

rule

Only applies to GGMs (including between-subjects networks) when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

binary

Logical. If TRUE then the weighted adjacency matrix will be converted into an unweighted adjacency matrix.

nodewise

Logical, only applies to GGMs (including between-subjects networks). If TRUE then the adjacency matrix will retain all coefficients in their original form. In this case, values in rows represent the coefficients predicting the columns.

d

Numeric. Only used for output of mlGVARsim , or simNet when lags = 1. Sets the number of decimal places to round the output to.

r

Numeric. Chooses which rows/columns to remove from the output, if desired.

avg

Logical. For netInts , determines whether to take the average two corresponding interaction terms.

empty

Logical. Determines the output of netInts when fit is not a moderated network. If TRUE then an empty list will be returned. If FALSE then a matrix of zeros will be returned.

mselect

Only used for netInts when there is more than one exogenous moderator. Allows the user to indicate which moderator should be used to construct the interaction matrix.

x <- fitNetwork(ggmDat, 'M')
net(x, threshold = .05)
netInts(x, threshold = TRUE)
y <- mlGVAR(mlgvarDat, 'M')
net(y, n = 'beta')
net(y, n = 'pcc')
net(y, n = 'between')
netInts(y)

ordinalize(
 data,
 m = NULL,
 nLevels = 5,
 thresholds = NULL,
 mthresh = NULL,
 mord = TRUE,
 minOrd = 3
)

data

An n x k dataframe or matrix containing only numeric values. Can also be a numeric vector.

m

The column number or name of the moderator variable, if applicable. Leave as NULL if there is no moderator, and set to TRUE if the moderator is the last column in the matrix or dataframe.

nLevels

Number of levels for the ordinal variables.

thresholds

List of length k, where each element is a numeric vector of length (nLevels - 1) containing the splitpoints for grouping each variable into ordered categories.

mthresh

Vector of length (nLevels - 1) containing thresholds to group values of the moderator into ordered categories.

mord

if FALSE, then the moderator will not be converted into an ordinal variable (if applicable).

minOrd

The minimum number of unique values allowed for each variable.

dat <- data.frame(sapply(1:5, function(z) rnorm(100)))
ord_dat <- ordinalize(dat)
# Including a moderator, without converting the moderator into an ordinal variable
ord_dat <- ordinalize(dat, m = 5, mord = FALSE)
colnames(dat)[5] <- 'M'
ord_dat <- ordinalize(dat, m = 'M', mord = FALSE)
# Use thresholds to break each variable into quartiles
thresh <- lapply(dat, function(z) quantile(z, probs = c(.25, .5, .75)))
ord_dat <- ordinalize(dat, thresholds = thresh)

## S3 method for class 'resample'
plot(x, what = "network", ...)

x

Output from the resample function.

what

Can be one of three options for all resample outputs: what = "network" will plot the final network model selected from resampling. what = "bootstrap" will run bootNet based on the final model to create bootstrapped estimates of confidence bands around each edge estimate. what = "coefs" will plot the confidence intervals based on the model parameters in the final network. Additionally, if the object was fit with sampMethod = "stability", a stability plot can be created with the "stability" option. Otherwise, if sampMethod = "bootstrap" or sampMethod = "split", a plot of the empirical distribution function of p-values can be displayed with the "pvals" option.

...

Additional arguments.

fit1 <- resample(ggmDat, m = 'M', niter = 10)
net(fit1)
netInts(fit1)
plot(fit1)
plot(fit1, what = 'coefs')
plot(fit1, what = 'bootstrap', multi = TRUE)
plot(fit1, what = 'pvals', outcome = 2, predictor = 4)
fit2 <- resample(gvarDat, m = 'M', niter = 10, lags = 1, sampMethod = 'stability')
plot(fit2, what = 'stability', outcome = 3)

plotBoot(
 x,
 type = "edges",
 net = "temporal",
 plot = "all",
 cor = 0.7,
 order = "mean",
 ci = 0.95,
 pairwise = TRUE,
 interactions = TRUE,
 labels = NULL,
 title = NULL,
 cis = "quantile",
 true = NULL,
 errbars = FALSE,
 vline = FALSE,
 threshold = FALSE,
 difference = FALSE,
 color = FALSE,
 text = FALSE,
 textPos = "value",
 multi = NULL,
 directedDiag = FALSE,
 ...
)
## S3 method for class 'bootNet'
plot(
 x,
 type = "edges",
 net = "temporal",
 plot = "all",
 cor = 0.7,
 order = "mean",
 ci = 0.95,
 pairwise = TRUE,
 interactions = TRUE,
 labels = NULL,
 title = NULL,
 cis = "quantile",
 true = NULL,
 errbars = FALSE,
 vline = FALSE,
 threshold = FALSE,
 difference = FALSE,
 color = FALSE,
 text = FALSE,
 textPos = "value",
 multi = NULL,
 directedDiag = FALSE,
 ...
)

x

Output from bootNet . Also some compatiblity with resample objects (when sampMethod != 'stability').

type

The outcome measure to plot. Options include: "edges", "strength", "ei", "outstrength", "instrength", "outei", "inei". The "out-" and "in-" options are only available for temporal networks. Moreover, both related options can be used together in temporal networks, by setting either type = c("outstrength", "instrength") or type = c("outei", "inei").

net

Determines which network to plot coefficients for. Options include: "ggm", "temporal", "contemporaneous", "between". Only relevant to SUR networks or mlGVAR objects.

plot

Primary use is to set as "none" or FALSE in order to return a table containing the constituents of the plot rather than create the plot itself. The options "all" and "both" each essentially indicate that both pairwise and interaction terms are plotted. Can also specify "pairwise" to only plot the pairwise terms, or "interactions" to only plot the interaction terms.

cor

Numeric value to indicate the correlation stability value to be plotted. Only applies to the case-drop bootstrapping output.

order

Determines how to arrange the predictors displayed in the plot. If TRUE, then defaults to "mean". If FALSE then defaults to "id". The "mean" option will arrange the values by the bootstrapped sample means. The "sample" option will arrange the values according to the statistics from the model fitted to the full sample. The "id" option will keep the variables in the same order that they appear in the dataframe. Not relevant to the case-drop bootstrap.

ci

Numeric value between 0 and 1 to specify the confidence level.

pairwise

Logical. Whether to plot pairwise relationships. Defaults to TRUE. If FALSE, this will override the "all" option of the plot argument.

interactions

Logical. Whether to plot interactions. Defaults to TRUE. If FALSE, this will override the "all" option of the plot argument. Only relevant to moderated networks.

labels

Logical. Determines whether to plot names of the predictors.

title

Character vector the title label.

cis

Either "quantile" or "se". If "quantile", then confidence bands will be computed based on quantiles (specified by the ci argument) of the bootstrapped resamples. If "se", then the confidence bands will be computed based on the standard errors associated with the sample statistics. Thus, the "se" argument will always produce a symmetric confidence band, whereas for "quantile" argument this is not necessary. Not relevant to outputs for the case-drop bootstrap.

true

Defaults to NULL, not relevant for the case-drop bootstrap. Can supply another output from fitNetwork , or an adjacency matrix, to serve as the true network in the plot. If there are interactions in the model, then a fitNetwork object is recommended. Alternatively, this argument can be extremely useful for simulated data – especially anything created with simNet . For whatever outcome (e.g., edges, strength, EI) is plotted, supplying another object to true will plot the values related to the true network, i.e., the data-generating model.

errbars

Logical. Not relevant to the case-drop bootstrap. If TRUE, then error bars are used rather than confidence bands. Can be useful to home in on specific variables and see their confidence interval.

vline

Logical or numeric. Not relevant to the case-drop bootstrap. If TRUE, then a dashed vertical line will be plotted at 0. If numeric, then the line will be plotted at the supplied intercept on the x-axis.

threshold

Numeric or logical. Not relevant to the case-drop bootstrap. Has a significant effect on the bootstrapped coefficient distributions. If TRUE, then the default p-value threshold is set to .05. A numeric value can specify a different threshold. Causes the bootNet function to run the object again, only to re-compute the bootstrapped distributions after applying a p-value threshold to the results of each model iteration. If NULL, all coefficient estimates are used in estimating the posterior distribution of each parameter.

difference

Logical. Not relevant to the case-drop bootstrap. If TRUE, then a difference plot is provided rather than a coefficient plot. In the difference plot, the diagonal squares reflect the fitted network coefficients for the the original sample. Black boxes indicate that the difference between the two edges, coefficients, or centrality values being compared is significantly different from 0. The significance level will have already been determined by the parameters used to fit the bootNet object. Gray boxes indicate the difference is not significantly different from 0.

color

Logical. Only applies when difference = TRUE. Determines whether to add colors that reflect the sign of the sample values. Reflected in the diagonal of the difference plot.

text

Logical. For difference plots, if TRUE then the statistics based on the full sample will be labeled in the diagonal boxes. For coefficient plots, setting this to TRUE will plot a label for each variable to reflect the proportion of times that it was selected across all bootstrapped iterations. Only relevant if a threshold was set for the fitted bootstrap models, either specified in the current function or was specified in creating the bootNet object. If a numeric value is provided, this will determine the size of the text label. Defaults to 1.2 when text = TRUE.

textPos

Supports the text argument for coefficient plots. Indicates the x-axis position of where to plot the coefficient labels. Generally will be numeric, but defaults to "value", which means that the text will be labeled on top each point on the plot.

multi

Useful when there are interactions in a model. If TRUE, the single plot with a facet for both pairwise and interaction terms is split into two separate plots. Allows for a more elegant side-by-side plot, and allows arguments that are restricted for plots of either pairwise or interactions (such as text) are plotted. This argument will eventually be expanded to allow one to plot combinations of edge and centrality plots.

directedDiag

See corresponding argument in the bootNet . function.

...

Additional arguments.

boot1 <- bootNet(ggmDat, caseDrop = TRUE)
plot(boot1)
plotBoot(boot1) # This functions the same as the command above
boot2 <- bootNet(ggmDat)
plot(boot2)
plot(boot2, difference = TRUE)

plotCoefs(
 fit,
 true = FALSE,
 alpha = 0.05,
 plot = TRUE,
 col = "blue",
 flip = TRUE,
 data = NULL,
 select = TRUE,
 size = 1,
 labels = TRUE,
 title = NULL,
 vars = "all"
)

fit

Output from fitNetwork , bootNet , or resample . Can also be the fixedNets or betweenNet elements of the mlGVAR output.

true

An adjacency matrix containing the true parameter values, if known. This can be used in conjunction with a simulated network, in that the user can supply the true network and plot those values against the estimated values.

alpha

Alpha level that is used to compute confidence intervals.

plot

Logical. If FALSE, a dataframe containing all of the confidence interval data will be returned.

col

Character string. Color of the points associated with the true values.

flip

Logical. If FALSE, the facets will be turned 90 degrees.

data

Supply the original dataset if not already included in the fit object.

select

Relevant to the resample output. Determines whether all variables should be plotted, or only those that were selected according to the resampling or variable selection procedure.

size

Numeric. Size of the point estimates.

labels

If logical, determines whether or not variable labels should be included. If a character vector, can be used to customize variable labels.

title

Custom plot title.

vars

Defaults to "all". Determines which variables should be plotted.

x <- fitNetwork(ggmDat)
plot(x, which.net = 'coefs')
plotCoefs(x) # This is the same as the above command

plotMods(
 nets,
 nodewise = FALSE,
 elsize = 2,
 vsize = NULL,
 elabs = TRUE,
 predict = NULL,
 layout = NULL,
 which.net = "temporal",
 ...
)

nets

List of network models fit with fitNetwork , where mval has been specified.

nodewise

See corresponding argument in plotNet .

elsize

Numeric value to indicate the size of the edge labels.

vsize

Numeric value to indicate the size of the nodes. If NULL, then a default value will be determined based on the number of nodes in the network.

elabs

If TRUE, then edges will be labeled with their numeric values.

predict

See corresponding argument in plotNet .

layout

Can be a character string, corresponding to the options in qgraph::qgraph , or can be a matrix that defines the layout (e.g., based on the qgraph::averageLayout function). Recommended to leave as NULL, so that the layout will be based on the list of networks provided.

which.net

See corresponding argument in plotNet .

...

Additional arguments.

data <- na.omit(psychTools::msq[, c('hostile', 'lonely', 'nervous', 'sleepy', 'depressed')])
fit0 <- fitNetwork(data, moderators = 'depressed', mval = 0)
fit1 <- fitNetwork(data, moderators = 'depressed', mval = 1)
fit2 <- fitNetwork(data, moderators = 'depressed', mval = 2)
fits <- list(fit0, fit1, fit2)
plotMods(fits)

plotNet(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'ggm'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'SURnet'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'mlGVAR'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'lmerVAR'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'ggmSim'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'mlGVARsim'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)
## S3 method for class 'GVARsim'
plot(
 x,
 which.net = "temporal",
 threshold = FALSE,
 layout = "spring",
 predict = FALSE,
 mnet = FALSE,
 names = TRUE,
 nodewise = FALSE,
 scale = FALSE,
 lag = NULL,
 con = "R2",
 cat = "nCC",
 covNet = FALSE,
 plot = TRUE,
 elabs = FALSE,
 elsize = 1,
 rule = "OR",
 binarize = FALSE,
 mlty = TRUE,
 mselect = NULL,
 ...
)

x

Output from any of the modnets model fitting or simulation functions.

which.net

When multiple networks exist for a single object, this allows the user to indicate which network to plot. For a GGM, all values of this argument return the same adjacency matrix. For a SUR network, "beta" and "temporal" plot the temporal network, while "pdc" plots the Partial Directed Correlations, or the standardized temporal network. "contemporaneous" and "pcc" plot the standardized contemporaneous network (Partial Contemporaneous Correlations). All of these terms apply for multilevel networks, but "between" can also plot the between-subjects network. Additionally, the value "coef" will plot the model coefficients and confidence intervals, defaulting to the plotCoefs function. Moreover, with GGMs or outputs from mlGVAR with a moderated between-subjects network, the value "ints" will call the intsPlot function. If a numeric or logical value is supplied, however, this argument will function as the threshold argument. A numeric value will set a threshold at the supplied value, while TRUE will set a threshold of .05.

threshold

A numeric or logical value to set a p-value threshold. TRUE will automatically set the threshold at .05.

layout

Character. Corresponds to the layout argument in the qgraph::qgraph function.

predict

If TRUE, then prediction error associated with each node will be plotted as a pie graph around the nodes. For continuous variables, the type of prediction error is determined by the con argument. For categorical variables, the type of error is determined by the cat argument. The desired value of con or can can be supplied directly into the present argument as well. Alternatively, another network model constituted by the same nodes can be supplied in order to plot the difference in prediction error, such as R-squared change.

mnet

Logical. If TRUE, the moderator will be plotted as a square "node" in the network, along with main effects represented as directed edges.

names

If TRUE, then the variable names associated with the model will be plotted as labels on the nodes. If FALSE, then nodes will be labeled with numbers rather than names. Alternatively, a character vector can be provided to serve as custom labels for the nodes.

nodewise

Only applies to GGMs. If TRUE, then nodewise edges will be plotted rather than the undirected averages of corresponding edges.

scale

Logical. Only applies when predict does not equal FALSE. The value of this argument is sent to the predictNet function. This argument will be removed.

lag

This argument will be removed. The function will automatically detect whether the network is based on time-lagged data.

con

Character string indicating which type of prediction error to plot for continuous variables, if predict does not equal FALSE. Options are: "R2", "adjR2", "MSE", "RMSE"

cat

Character string indicating which type of prediction error to plot for categorical variables, if predict does not equal FALSE. Options are: "nCC", "CC", "CCmarg"

covNet

Logical. Only applies when a covariate is modeled. Allows the covariate to be plotted as a separate square "node".

plot

Logical. If FALSE, then a qgraph object will be returned rather than plotted.

elabs

Logical. If TRUE, the values of the edges will be plotted as labels on the edges.

elsize

numeric

rule

Only applies to GGMs (including between-subjects networks) when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

binarize

Logical. If TRUE, the network will be plotted as an unweighted network. Only applies to GGMs.

mlty

Logical. If FALSE, then moderated edges are displayed as solid lines. If TRUE, then moderated edges are shown as dashed lines.

mselect

If the model contains more than one moderator, input the character string naming which moderator you would like the plot to reflect. Only affects which lines are dashed or solid. Not compatible with the mnet argument.

...

Additional arguments.

fit1 <- fitNetwork(ggmDat)
plot(fit1)
plotNet(fit1) # This and the command above produce the same result
fit2 <- fitNetwork(gvarDat, moderators = 'M', lags = 1)
plot(fit2, 'pdc') # Partial Directed Correlations
plot(fit2, 'pcc') # Partial Contemporaneous Correlations

plotNet2(
 object,
 whichNets = NULL,
 whichTemp = c("temporal", "PDC"),
 titles = c("PDC ", "PCC "),
 ...
)

object

Output from fitNetwork , specifically with a SUR model.

whichNets

Vector of length 2 indicating which networks to plot. "beta" and "temporal" both refer to the unstandardized temporal network coefficients, while "PDC" refers to the standardized temporal network. "PCC" and "contemporaneous" both refer to the standardized residual covariance matrix (the contemporaneous network). If the object is fitted with mlGVAR or lmerVAR , then "between" is also an option for plotting the between-subjects network.

whichTemp

Which version of the temporal network should be plotted, either "temporal" or "PDC". This argument is ignored if whichNets is not NULL.

titles

Character vector of length 2 where custom names for the two network plots can be supplied.

...

Additional arguments.

x <- fitNetwork(gvarDat, lags = TRUE)
plotNet2(x)

plotNet3(
 object,
 ...,
 nets = c("temporal", "contemporaneous", "between"),
 titles = TRUE,
 l = 3,
 label = NULL,
 xpos = 0,
 ypos = 0.5
)

object

Output from mlGVAR or lmerVAR .

...

Additional arguments.

nets

Character vector of length 3 indicating which networks to plot, and in what order. Same options as for which.net in plotNet .

titles

If TRUE, then uses default titles for temporal, contemporaneous, and between-subject networks. If FALSE, then no titles will be used. Can also be a character vector to provide custom plot titles.

l

A numeric value to indicate a type of pane layout.

label

Can include a character string for text annotation.

xpos

Numeric, x-axis value for where the text annotation should go. Between 0 and 1.

ypos

numeric, y-axis value for where the text annotation should go. Between 0 and 1.

x <- mlGVAR(mlgvarDat, 'M')
plotNet3(x)

plotPower(
 x,
 by = "type",
 yvar = "default",
 yadd = NULL,
 hline = 0.8,
 xlab = "Number of cases",
 title = NULL,
 ...
)
## S3 method for class 'mnetPower'
plot(
 x,
 by = "type",
 yvar = "default",
 yadd = NULL,
 hline = 0.8,
 xlab = "Number of cases",
 title = NULL,
 ...
)

x

mnetPowerSim output

by

In development. Currently only supports "type" for creating different facets for Pairwise and Interaction effects. "network" for creating facets based on different networks (e.g., temporal, contemporaneous). "p" for creating facets based on the number of nodes in the network.

yvar

The performance metrics to plot. Options include: "sensitivity", "specificity", "correlation", "precision", "MAE", "FDR", "accuracy". The option "default" automatically sets this to sensitivity, specificity, and correlation.

yadd

Specify additional performance metrics to plot. The final performance metrics that end up being plotted are simply: c(yvar, yadd). Thus, this argument is only useful as a shortcut for keeping the default values of yvar, but adding more metrics to plot.

hline

Numeric value between 0 and 1 for where to plot a horizontal line of interest. Can set to FALSE to remove line.

xlab

Character string for the x-axis label.

title

Character string for the title of the plot.

...

Additional arguments.

x <- mnetPowerSim(niter = 10, N = c(100, 200))
summary(x)
plot(x)

plotPvals(x, outcome = 1, predictor = 1, title = TRUE, alpha = 0.05)

x

Output from resample , given that sampMethod = "bootstrap" or sampMethod = "split".

outcome

Character string or numeric value (in terms of columns in the dataset) to indicate which outcome to plot the p-value distribution for.

predictor

Character string or numeric value (in terms of columns in the dataset) to indicate which predictor to plot the p-value distribution for.

title

If TRUE, then a default title will be given according to the outcome and predictor that are specified. If FALSE, then no title will be plotted. A custom title may also be supplied by the user.

alpha

The false discovery rate. Defaults to .05

x <- resample(ggmDat, sampMethod = "bootstrap")
plot(x, what = 'pvals')
plot(x, 'pvals', outcome = 'V2', predictor = 'V1')

plotStability(
 x,
 outcome = 1,
 s = c("simult", "split1", "split2"),
 thresh = 0.5,
 typeLegend = TRUE
)

x

Output of resample where sampMethod = "stability".

outcome

Character string or numeric value (in terms of columns in the dataset) to indicate which outcome to plot the stability path for.

s

Character string or numeric value. This indicates which stability path to return a plot for. Either the first sample split "split1", the second sample split "split2", or the path for simultaneous selection "simult", which is the default.

thresh

The selection threshold, which is represented as a horizontal red line on the plot. Defaults to .5

typeLegend

Logical. If FALSE, linetype legend is removed. Only applies if there is a moderator in the model.

x <- resample(ggmDat, sampMethod = "stability")
plot(x, what = "stability")
plot(x, 'stability', outcome = 'V3')

predictNet(object, data = NULL, all = FALSE, scale = FALSE)

object

Output from fitNetwork or mlGVAR . If using output from mlGVAR , then one of the two networks must be provided (i.e., either fixedNets or betweenNet).

data

The dataset used to fit the network model, or another network of the same type and size to be compared with the network specified in the first argument. If the prediction error for only one network is desired, and the dataset is included as an element of the relevant object, then this can be left as NULL.

all

if TRUE then returns a list containing the observed outcomes used to fit the models, their predicted values, and the prediction error for each outcome.

scale

Logical; determines whether or not to standardize the data before computing prediction error. This argument will be removed.

fit1 <- fitNetwork(ggmDat, covariates = 'M')
fit2 <- fitNetwork(ggmDat, moderators = 'M')
predictNet(fit1)
predictNet(fit1, all = TRUE)
predictNet(fit2, fit1) # Find the differences in prediction error across the two models

resample(
 data,
 m = NULL,
 niter = 10,
 sampMethod = "bootstrap",
 criterion = "AIC",
 method = "glmnet",
 rule = "OR",
 gamma = 0.5,
 nfolds = 10,
 nlam = 50,
 which.lam = "min",
 threshold = FALSE,
 bonf = FALSE,
 alpha = 0.05,
 exogenous = TRUE,
 split = 0.5,
 center = TRUE,
 scale = FALSE,
 varSeed = NULL,
 seed = NULL,
 verbose = TRUE,
 lags = NULL,
 binary = NULL,
 type = "g",
 saveMods = TRUE,
 saveData = FALSE,
 saveVars = FALSE,
 fitit = TRUE,
 nCores = 1,
 cluster = "mclapply",
 block = FALSE,
 beepno = NULL,
 dayno = NULL,
 ...
)

data

n x k dataframe. Cannot supply a matrix as input.

m

Character vector or numeric vector indicating the moderator(s), if any. Can also specify "all" to make every variable serve as a moderator, or 0 to indicate that there are no moderators. If the length of m is k - 1 or longer, then it will not be possible to have the moderators as exogenous variables. Thus, exogenous will automatically become FALSE.

niter

Number of iterations for the resampling procedure.

sampMethod

Character string indicating which type of procedure to use. "bootstrap" is a standard bootstrapping procedure. "split" is the multi-sample split procedure where the data are split into disjoint training and test sets, the variables to be modeled are selected based on the training set, and then the final model is fit to the test set. "stability" is stability selection, where models are fit to each of two disjoint subsamples of the data, and it is calculated how frequently each variable is selected in each subset, as well how frequently they are simultaneously selected in both subsets at each iteration.

criterion

The criterion for the variable selection procedure. Options include: "cv", "aic", "bic", "ebic", "cp", "rss", "adjr2", "rsq", "r2". "CV" refers to cross-validation, the information criteria are "AIC", "BIC", "EBIC", and "Cp", which refers to Mallow's Cp. "RSS" is the residual sum of squares, "adjR2" is adjusted R-squared, and "Rsq" or "R2" is R-squared. Capitalization is ignored. For methods based on the LASSO, only "CV", "AIC", "BIC", "EBIC" are available. For methods based on subset selection, only "Cp", "BIC", "RSS", "adjR2", "R2" are available.

method

Character string to indicate which method to use for variable selection. Options include "lasso" and "glmnet", both of which use the LASSO via the glmnet package (either with glmnet::glmnet or glmnet::cv.glmnet , depending upon the criterion). "subset", "backward", "forward", "seqrep", all call different types of subset selection using the leaps::regsubsets function. Finally "glinternet" is used for applying the hierarchical lasso, and is the only method available for moderated network estimation (either with glinternet::glinternet or glinternet::glinternet.cv , depending upon the criterion). If one or more moderators are specified, then method will automatically default to "glinternet".

rule

Only applies to GGMs (including between-subjects networks) when a threshold is supplied. The "AND" rule will only preserve edges when both corresponding coefficients have p-values below the threshold, while the "OR" rule will preserve an edge so long as one of the two coefficients have a p-value below the supplied threshold.

gamma

Numeric value of the hyperparameter for the "EBIC" criterion. Only relevant if criterion = "EBIC". Recommended to use a value between 0 and .5, where larger values impose a larger penalty on the criterion.

nfolds

Only relevant if criterion = "CV". Determines the number of folds to use in cross-validation.

nlam

if method = "glinternet", determines the number of lambda values to evaluate in the selection path.

which.lam

Character string. Only applies if criterion = "CV". Options include "min", which uses the lambda value that minimizes the objective function, or "1se" which uses the lambda value at 1 standard error above the value that minimizes the objective function.

threshold

Logical or numeric. If TRUE, then a default value of .05 will be set. Indicates whether a threshold should be placed on the models at each iteration of the sampling. A significant choice by the researcher.

bonf

Logical. Determines whether to apply a bonferroni adjustment on the distribution of p-values for each coefficient.

alpha

Type 1 error rate. Defaults to .05.

exogenous

Logical. Indicates whether moderator variables should be treated as exogenous or not. If they are exogenous, they will not be modeled as outcomes/nodes in the network. If the number of moderators reaches k - 1 or k, then exogenous will automatically be FALSE.

split

If sampMethod == "split" or sampMethod = "stability" then this is a value between 0 and 1 that indicates the proportion of the sample to be used for the training set. When sampMethod = "stability" there isn't an important distinction between the labels "training" and "test", although this value will still cause the two samples to be taken of complementary size.

center

Logical. Determines whether to mean-center the variables.

scale

Logical. Determines whether to standardize the variables.

varSeed

Numeric value providing a seed to be set at the beginning of the selection procedure. Recommended for reproducible results. Importantly, this seed will be used for the variable selection models at each iteration of the resampler. Caution this means that while each model is run with a different sample, it will always have the same seed.

seed

Can be a single value, to set a seed before drawing random seeds of length niter to be used across iterations. Alternatively, one can supply a vector of seeds of length niter. It is recommended to use this argument for reproducibility over the varSeed argument.

verbose

Logical. Determines whether information about the modeling progress should be displayed in the console.

lags

Numeric or logical. Can only be 0, 1 or TRUE or FALSE. NULL is interpreted as FALSE. Indicates whether to fit a time-lagged network or a GGM.

binary

Numeric vector indicating which columns of the data contain binary variables.

type

Determines whether to use gaussian models "g" or binomial models "c". Can also just use "gaussian" or "binomial". Moreover, a vector of length k can be provided such that a value is given to every variable. Ultimately this is not necessary, though, as such values are automatically detected.

saveMods

Logical. Indicates whether to save the models fit to the samples at each iteration or not.

saveData

Logical. Determines whether to save the data from each subsample across iterations or not.

saveVars

Logical. Determines whether to save the variable selection models at each iteration.

fitit

Logical. Determines whether to fit the final selected model on the original sample. If FALSE, then this can still be done with fitNetwork and modSelect .

nCores

Numeric value indicating the number of CPU cores to use for the resampling. If TRUE, then the parallel::detectCores function will be used to maximize the number of cores available.

cluster

Character vector indicating which type of parallelization to use, if nCores > 1. Options include "mclapply" and "SOCK".

block

Logical or numeric. If specified, then this indicates that lags != 0 or lags != NULL. If numeric, then this indicates that block bootstrapping will be used, and the value specifies the block size. If TRUE then an appropriate block size will be estimated automatically.

beepno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with dayno argument.

dayno

Character string or numeric value to indiciate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with beepno argument.

...

Additional arguments.

fit1 <- resample(ggmDat, m = 'M', niter = 10)
net(fit1)
netInts(fit1)
plot(fit1)
plot(fit1, what = 'coefs')
plot(fit1, what = 'bootstrap', multi = TRUE)
plot(fit1, what = 'pvals', outcome = 2, predictor = 4)
fit2 <- resample(gvarDat, m = 'M', niter = 10, lags = 1, sampMethod = 'stability')
plot(fit2, what = 'stability', outcome = 3)

sampleSize(p, m = 0, lags = 0, print = TRUE)

p

Number of nodes

m

Number of moderator variables (defaults to 0)

lags

Number of lags (currently only supports 0 and 1)

print

if FALSE, then the minimum sample size is returned and can be assigned to an object.

sampleSize(p = 10)
sampleSize(p = 10, m = 1)
sampleSize(p = 10, m = 1, lags = 1)
minSamp <- sampleSize(p = 10, m = 1, lags = 1, print = FALSE)

selected(object, threshold = FALSE, mod = c("temporal", "between"))

object

Output from either fitNetwork or mlGVAR

threshold

Can be a numeric value between 0 and 1, or defaults to .05 when set to TRUE

mod

Only relevant to models fit with mlGVAR

fit1 <- fitNetwork(ggmDat)
selected(fit1)
fit2 <- mlGVAR(mlgvarDat, m = 'M', verbose = FALSE)
selected(fit2, threshold = TRUE, mod = 'temporal') # Can also set to 'between'
fit3 <- fitNetwork(gvarDat, moderators = 'M', type = 'varSelect', lags = 1)
selected(fit3)

simNet(
 N = 100,
 p = 5,
 m = FALSE,
 m2 = 0.1,
 b1 = NULL,
 b2 = NULL,
 sparsity = 0.5,
 intercepts = NULL,
 nIter = 250,
 msym = FALSE,
 onlyDat = FALSE,
 pbar = TRUE,
 div = 10,
 gibbs = TRUE,
 ordinal = FALSE,
 nLevels = 5,
 mord = FALSE,
 time = TRUE,
 mbinary = FALSE,
 minOrd = 3,
 m1 = NULL,
 m1_range = NULL,
 m2_range = c(0.1, 0.3),
 modType = "none",
 lags = NULL,
 V = 2,
 skewErr = FALSE,
 onlyNets = FALSE,
 netArgs = NULL,
 nCores = 1,
 cluster = "SOCK",
 getChains = FALSE,
 const = 1.5,
 fixedPar = NULL,
 V2 = 1,
 ...
)

N

Numeric value. Total number of subjects.

p

Numeric value. Total number of nodes (excluding moderator).

m

If a value is provided, a moderator is generated and named M in the resultant data. If TRUE, then a normal distribution with a mean of 0 will be used to generate the initial value of m, which will serve as the population mean for m throughout the simulation. If a numeric value is provided, then this will serve as the population mean, and all subsequent draws will be taken from a normal distribution with that mean. If m = "binary", then this will simply set the argument mbinary = TRUE. If m = "ordinal", this will set mord = TRUE. To simulate m from a skewed distribution, there are two options: if m = "skewed", then the alpha parameter of the sn::rmsn will automatically be set to 3. Alternatively, a vector of length two can be supplied, containing the element "skewed" as well as the desired value of alpha. Lastly, a function can be provided for m if the user wishes to sample m from another distribution. The requirement is that the function have only one argument, and only returns a single numeric value. The input of the argument should be the location parameter of the desired sampling distribution.

m2

Numeric. If m2 >= 1, then this will determine the number of interaction effects between the moderator and some node in the network. If a value between 0 and 1 is provided, then this determines the probability of any given edge being moderated by the moderator.

b1

Can provide an adjacency matrix to use for generating data.

b2

Can provide an interaction matrix for generated moderated data.

sparsity

Numeric value between 0 and 1. Determines the sparsity of sampled network matrices.

intercepts

A vector of means for sampling node values.

nIter

Number of iterations for generating each instance of a datapoint with the Gibbs sampler.

msym

If TRUE then will force the interaction matrix to be symmetric.

onlyDat

If TRUE then the function only returns the simulated data.

pbar

If TRUE then a progress bar will be shown as samples are generated.

div

A value to use as a sign that the sampler diverged. Can be increased based on expected range of values. If a datapoint is larger than div, then the sampler will stop.

gibbs

If TRUE, then Gibbs sampling will be used. Otherwise, data are generated from the mvtnorm::rmvnorm function based on the partial correlation matrix that is created.

ordinal

Logical. Determines whether to generate ordinal values or not.

nLevels

Number of levels for the ordinal variables. Only relevant if ordinal is not FALSE.

mord

Logical. Determines whether the moderator variable should be simulated as ordinal.

time

If TRUE then the time it takes to simulate the data is printed to screen at the end of the sampling.

mbinary

Logical. Determines whether the moderator should be a binary variable.

minOrd

The minimum number of unique values allowed for each variable.

m1

Functions similarly to m2, except that this argument refers to the number/probability of main effects of the moderator on any given node.

m1_range

Numeric vector of length 2. The range of values for moderator main effect coefficients.

m2_range

Numeric vector of length 2. The range of values for moderator interaction effect coefficients.

modType

Determines the type of moderation to employ, such as "none", "full", "partial". If modType = "full", then for any interaction terms there will be full moderation, such that all pairwise relationships for moderated paths will be set to zero. If modType = "partial", then pairwise edges for moderated paths will always be nonzero. If modType = "none", no constraints will be applied (e.g., could produce a mix between full and partial moderation).

lags

If TRUE or 1, then arguments are rerouted to the mlGVARsim function to simulate temporal data for a single individual.

V

Numeric, either 1 or 2. Determines whether to randomize the order of simulating node values at each iteration of the Gibbs sampler. If V = 2, then the order is randomized at each iteration. If V = 1, then the sampler moves through the nodes from the first to the last in order at each iteration.

skewErr

The skewness parameter for the alpha argument in the sn::rmsn function. Only relevant when gibbs = FALSE and no moderator is specified.

onlyNets

If TRUE then only the network models are returned, without the data. Could be used to create random models and then simulate data by another method.

netArgs

Only for use by the internal function modnets:::simNet2, which serves as a wrapper for the current function to prevent it from failing.

nCores

Numeric value indicating the number of CPU cores to use for the resampling. If TRUE, then the parallel::detectCores function will be used to maximize the number of cores available.

cluster

Character vector indicating which type of parallelization to use, if nCores > 1. Options include "mclapply" and "SOCK".

getChains

Logical. Determines whether to return the data-generating chains from the Gibbs sampler.

const

Numeric. The constant to be used by the internal modnets:::simPcor function.

fixedPar

Numeric. If provided, then this will be set as the coefficient value for all edges in the network. Provides a way to standardize the parameter values while varying the sparsity of the network. If length(fixedPar) == 1, then the same value will be used for all parameters. If length(fixedPar) == 2, then the first value will be for pairwise relationships, and the second value will be for interaction terms.

V2

If V2 = 1 and m2 is between 0 and 1, the number of interaction terms in the model will be determined by multiplying m2 with the number of elements in the interaction matrix and taking the ceiling.

...

Additional arguments.

# Generate a moderated GGM along with data
set.seed(1)
x <- simNet(N = 100, p = 3, m = TRUE)
net(x) # Get data-generating adjacency matrix
netInts(x) # Get data-generating interaction matrix
plot(x) # Plot the moderated network that generated the data
# Generate a single-subject GVAR model with data
set.seed(1)
x <- simNet(N = 500, p = 3, m = TRUE, lags = 1)
net(x, n = 'temporal') # Get the data-generating time-lagged adjacency matrix
net(x, n = 'contemporaneous') # Get the data-generating standardized residual covariance matrix
plot(x, which.net = 'beta') # 'beta' is another way of referring to the temporal network
plot(x, which.net = 'pcc') # 'pcc' is another way of referring to the contemporaneous network

## S3 method for class 'mnetPower'
summary(object, ind = "all", order = NULL, decreasing = FALSE, ...)

object

Output from mnetPowerSim function.

ind

Character string or vector to indicate which aspects of the results to view. If "means", then only the means will be returned for all performance indices. "sds" returns the standard deviations, "ses" returns the standard errors, and "medians" returns the medians. These statistics describe the sample distributions according to each combination of input parameters, and with regard to all performance indices. Any combination of these options will return a list with each table as a separate element. "all" returns a list of length 4 with tables for all 4 types of statistic.

order

Character string referring to which output index to organize output by.

decreasing

Logical. Determines whether to organize values from highest to lowest or vice versa according to the value of the order argument.

...

Additional arguments.

x <- mnetPowerSim(niter = 10, N = c(100, 200))
summary(x)
plot(x)

varSelect(
 data,
 m = NULL,
 criterion = "AIC",
 method = "glmnet",
 lags = NULL,
 exogenous = TRUE,
 type = "g",
 center = TRUE,
 scale = FALSE,
 gamma = 0.5,
 nfolds = 10,
 varSeed = NULL,
 useSE = TRUE,
 nlam = NULL,
 covs = NULL,
 verbose = TRUE,
 beepno = NULL,
 dayno = NULL
)

data

n x k dataframe or matrix.

m

Character vector or numeric vector indicating the moderator(s), if any. Can also specify "all" to make every variable serve as a moderator, or 0 to indicate that there are no moderators. If the length of m is k - 1 or longer, then it will not be possible to have the moderators as exogenous variables. Thus, exogenous will automatically become FALSE.

criterion

The criterion for the variable selection procedure. Options include: "cv", "aic", "bic", "ebic", "cp", "rss", "adjr2", "rsq", "r2". "CV" refers to cross-validation, the information criteria are "AIC", "BIC", "EBIC", and "Cp", which refers to Mallow's Cp. "RSS" is the residual sum of squares, "adjR2" is adjusted R-squared, and "Rsq" or "R2" is R-squared. Capitalization is ignored. For methods based on the LASSO, only "CV", "AIC", "BIC", "EBIC" are available. For methods based on subset selection, only "Cp", "BIC", "RSS", "adjR2", "R2" are available.

method

Character string to indicate which method to use for variable selection. Options include "lasso" and "glmnet", both of which use the LASSO via the glmnet package (either with glmnet::glmnet or glmnet::cv.glmnet , depending upon the criterion). "subset", "backward", "forward", "seqrep", all call different types of subset selection using the leaps::regsubsets function. Finally "glinternet" is used for applying the hierarchical lasso, and is the only method available for moderated network estimation (either with glinternet::glinternet or glinternet::glinternet.cv , depending upon the criterion). If one or more moderators are specified, then method will automatically default to "glinternet".

lags

Numeric or logical. Can only be 0, 1 or TRUE or FALSE. NULL is interpreted as FALSE. Indicates whether to fit a time-lagged network or a GGM.

exogenous

Logical. Indicates whether moderator variables should be treated as exogenous or not. If they are exogenous, they will not be modeled as outcomes/nodes in the network. If the number of moderators reaches k - 1 or k, then exogenous will automatically be FALSE.

type

Determines whether to use gaussian models "g" or binomial models "c". Can also just use "gaussian" or "binomial". Moreover, a vector of length k can be provided such that a value is given to every variable. Ultimately this is not necessary, though, as such values are automatically detected.

center

Logical. Determines whether to mean-center the variables.

scale

Logical. Determines whether to standardize the variables.

gamma

Numeric value of the hyperparameter for the "EBIC" criterion. Only relevant if criterion = "EBIC". Recommended to use a value between 0 and .5, where larger values impose a larger penalty on the criterion.

nfolds

Only relevant if criterion = "CV". Determines the number of folds to use in cross-validation.

varSeed

Numeric value providing a seed to be set at the beginning of the selection procedure. Recommended for reproducible results.

useSE

Logical. Only relevant if method = "glinternet" and criterion = "CV". Indicates whether to use the standard error of the estimates across folds, if TRUE, or to use the standard deviation, if FALSE.

nlam

if method = "glinternet", determines the number of lambda values to evaluate in the selection path.

covs

Numeric or character string indicating a variable to be used as a covariate. Currently not working properly.

verbose

Logical. Determines whether to provide output to the console about the status of the procedure.

beepno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with dayno argument.

dayno

Character string or numeric value to indicate which variable (if any) encodes the survey number within a single day. Must be used in conjunction with beepno argument.

vars1 <- varSelect(ggmDat, criterion = 'BIC', method = 'subset')
fit1 <- fitNetwork(ggmDat, type = vars1)
vars2 <- varSelect(ggmDat, criterion = 'CV', method = 'glmnet')
fit2 <- fitNetwork(ggmDat, type = vars2, which.lam = 'min')
# Add a moderator
vars3 <- varSelect(ggmDat, m = 'M', criterion = 'EBIC', gamma = .5)
fit3 <- fitNetwork(ggmDat, moderators = 'M', type = vars3)