Use cartographic variables to classify forest categories. Balanced Mulit-Class Classification problem. Exploratory Data Analysis performed to identify promising features, visualising data interactions. Feature selection to reduce the number of features from 54 to 12. Performed predictive modelling using tree based methods Random Forest and XGBoost. Kaggle accuracy of 76.4%.
-
Notifications
You must be signed in to change notification settings - Fork 0
DnanaDev/Forest_Cover_Analysis_Prediction
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Latest commit | ||||
Repository files navigation
About
EDA and Predicitve Modelling on UCI forest cover dataset, kaggle challenge public score 0.764.
Resources
Stars
Watchers
Forks
Releases
No releases published