Permutation Importance

Our A.I.4trading team explains the importance of feature analysis in machine learning and deep learning models through a series of three report. In this first paper, we analyze a technique known as “permutation importance” which will be applied to Random forest and Extremely randomized Trees models. We also explain the theoretical basis of this technique and its strengths compared to feature importance evaluation through the “Gini index” for tree models and we provide experimental results of this technique.