10.2 Filter-Based Feature Selection
You may remember from Chapter 3.1 that we trained many, many linear regression models with various combinations of independent variables. For example, we'd test one model, discover a few insignficant variables, remove those variables, add new variables, test the new model, rinse and repeat and repeat and repeat. Wouldn't it be nice if a smart tool could help you speed up that process? That's exactly the purpose of the filter-based feature selection and permutation feature importance tools in ML Studio. Follow along with the videos below to learn how these tools work: