This tutorial shows how to do machine learning in R using the caret package. It covers the basics of data partitioning, pre-processing, model training, and hyperparameter tuning using the unified framework of the caret package.
00:00 Introduction
02:12 Data setup for a binary classification problem
04:00 Data partition into training and test sets
05:54 Pre-processing of data
09:50 Introduction to train() function
11:50 Setting up resampling method using trainControl()
12:30 Using train() to fit a model
14:33 Changing the metric for choosing optimal parameter values
16:40 Assesing classification accuracy through the confusion matrix
19:40 Hyperparameter Tuning
20:31 Aside on Elastic-Net type logistical regression classification model
22:07 Changing the model call in train()
24:50 Obtaining variable contribution using varImp()
26:34 Conclusion