Orthogonal variance decomposition based feature selection

TL;DR

This paper introduces an orthogonal variance decomposition method for feature selection that effectively accounts for feature interactions, leading to more accurate identification of relevant features and improved model performance.

Contribution

It presents a novel feature selection approach using orthogonal variance decomposition to directly measure feature contributions considering interactions.

Findings

Accurately identifies relevant features

Improves numerical model accuracy

Efficiently evaluates feature contributions

Abstract

Existing feature selection methods fail to properly account for interactions between features when evaluating feature subsets. In this paper, we attempt to remedy this issue by using orthogonal variance decomposition to evaluate features. The orthogonality of the decomposition allows us to directly calculate the total contribution of a feature to the output variance. Thus we obtain an efficient algorithm for feature evaluation which takes into account interactions among features. Numerical experiments demonstrate that our method accurately identifies relevant features and improves the accuracy of numerical models.