Python implementation of Sequential Forward Feature Selection from scratch.

• The program will take one input: a dataset where the last column is the class variable.
• The program will load the dataset and then use the wrapper approach with a sequential forward selection strategy to find a set of essential features.
• Stratified 5-fold cross-validation was used for measuring accuracy.
• The program will keep adding the features as long as there is some improvement in the classification accuracy.
• The output of the program will be the set of important features on the console.