In this video, we explain the step-forward feature selection algorithm using a practical example. Step forward feature selection is a greedy algorithm that evaluates feature subsets by adding a feature at a time and measuring the change in performance of the machine learning model. It is greedy, because it tests all possible additions of a single feature to a subset. It belongs to the group of wrapper methods.
In the video we discuss how it works, how models are built by adding one feature at a time, and how to determine the optimal number of features for your model. We also introduce MLXtend, a Python open source library with an efficient implementation of this feature selection process.
Want to learn more? Check out our course: https://www.trainindata.com/p/feature-selection-for-machine-learning
