April 2018
Beginner to intermediate
282 pages
6h 52m
English
Recursive feature elimination is based on the idea of recursively constructing a model by removing the features, building the model with the remaining features, and computing the model's accuracy. This process is repeated until all features in the dataset are exhausted. It is a greedy optimization method to find the best performing subset of features and then rank them according to when they were eliminated.
In the following example code, the HR attrition dataset is used to illustrate the use of recursive feature elimination (RFE). The stability of the RFE method is heavily dependent on the type of algorithm used. For our demonstration, we have used the LogisticRegression method:
#Recursive Feature Elimination ...