Machine Learning with R Cookbook - Second Edition

How it works...

In this recipe, we present how to use the FSelector package to select the most influential features. We first demonstrate how to use the feature ranking approach. In the feature ranking approach, the algorithm first employs a weight function to generate weights for each feature. Here, we use the random forest algorithm with the mean decrease in accuracy (where importance.type = 1) as the importance measurement to gain the weights of each attribute. Besides the random forest algorithm, you can select other feature ranking algorithms (for example, chi.squared, information.gain) from the FSelector package. Then, the process sorts attributes by their weight. Finally, we can obtain the top five features from the sorted feature ...