Applied Predictive Modeling with caret
Create synthetic data using twoClassSim
Quickly explore the data using skimr and xray
Split the dataset into train/test with an index
Setting up train control in caret
Cross-validation method and settings
Subsampling to deal with class-imbalance (mentioned but not implemented)
Placeholder regression example
Classification example
Logistic Regression (glm), Elastic Net (glmnet), Random Forest (ranger)
Using summary, variable importance, plot on fit object
Prediction on unseen data: class; class probability
In-sample: ROC, Sensitivity (true positive rate), Specificity (true negative rate)
Confusion matrices
Model dissimilarity using Jaccard distance
Linear ensembles
Meta-Model ensembles
Recursive Feature Elimination
Simulated Annealing
Genetic Algorithm