The random forest functions create a predictive model based on a combination of the Classification and Regression Trees (CART) algorithm for training decision trees and the ensemble learning method of bagging.
The random forest functions are:
- Forest_Drive, which builds a predictive model based on training data.
- Forest_Predict, which uses the model generated by the Forest_Drive function to analyze the input data and make predictions.
- Forest_Analyze, which analyzes the model generated by the Forest_Drive function and gives weights to the variables used in the model. This helps you understand the basis by which the Forest_Predict function makes predictions.
You can use the Forest_Drive and Forest_Predict functions to generate predictions input for the Receiver Operating Characteristic (ROC) function.