Commit e8a80e1

authored

Merge pull request #456 from voberoi/autocorr-fix

Ch. 3 correction: MCMC *not* exploring the space well likely exhibits high autocorr.

2 parents e575439 + 7a5dd0e commit e8a80e1Copy full SHA for e8a80e1

File tree

-3

lines changed

-3

lines changed

Lines changed: 1 addition & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -1091,7 +1091,7 @@`
`1091`	`1091`	`"\n",`
`1092`	`1092`	"By the nature of the MCMC algorithm, we will always be returned samples that exhibit autocorrelation (this is because of the step `from your current position, move to a position near you`).\n",
`1093`	`1093`	`"\n",`
`1094`		`- "A chain that is [Isn't meandering exploring?] exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n",`
	`1094`	`+ "A chain that is not exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n",`
`1095`	`1095`	`"\n",`
`1096`	`1096`	"This does not imply that a converged MCMC has low autocorrelation. Hence low autocorrelation is not necessary for convergence, but it is sufficient. PyMC has a built-in autocorrelation plotting function in the `Matplot` module. "
`1097`	`1097`	`]`

Lines changed: 1 addition & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -1073,7 +1073,7 @@`
`1073`	`1073`	`"\n",`
`1074`	`1074`	"By the nature of the MCMC algorithm, we will always be returned samples that exhibit autocorrelation (this is because of the step `from your current position, move to a position near you`).\n",
`1075`	`1075`	`"\n",`
`1076`		`- "A chain that is [Isn't meandering exploring?] exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n",`
	`1076`	`+ "A chain that is not exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n",`
`1077`	`1077`	`"\n",`
`1078`	`1078`	"This does not imply that a converged MCMC has low autocorrelation. Hence low autocorrelation is not necessary for convergence, but it is sufficient. PyMC3 has a built-in autocorrelation plotting function in the `plots` module. "
`1079`	`1079`	`]`

Lines changed: 1 addition & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -1909,7 +1909,7 @@`
`1909`	`1909`	`"\n",`
`1910`	`1910`	"By the nature of the MCMC algorithm, we will always be returned samples that exhibit autocorrelation (this is because of the step `from your current position, move to a position near you`).\n",
`1911`	`1911`	`"\n",`
`1912`		`- "A chain that is [Isn't meandering exploring?] exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n",`
	`1912`	`+ "A chain that is not exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n",`
`1913`	`1913`	`"\n",`
`1914`	`1914`	`"This does not imply that a converged MCMC has low autocorrelation. Hence low autocorrelation is not necessary for convergence, but it is sufficient. TFP has a built-in autocorrelation tools as well. "`
`1915`	`1915`	`]`

Comments

(0)