Feature Selection via Coalitional Game Theory

By Shay B. Cohen, Gideon Dror, Eytan Ruppin. 2007

153 citations recorded by the source corpus.

We present and study the contribution-selection algorithm (CSA), a novel algorithm for feature selection. The algorithm is based on the multiperturbation shapley analysis (MSA), a framework that relies on game theory to estimate usefulness. The algorithm iteratively estimates the usefulness of features and selects them accordingly, using either forward selection or backward elimination. It can optimize various performance measures over unseen data such as accuracy, balanced error rate, and area under receiver-operator-characteristic curve. Empirical comparison with several other existing feature selection methods shows that the backward elimination variant of CSA leads to the most accurate classification results on an array of data sets.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Systems

Citations in the corpus

Visualizing the Feature Importance for Black Box Models (2018, methodology): “roximation of the model performance from Eq. (6). We therefore suggest an ecient procedure in Algorithm 2. The Shapley value satises the following four desirable properties as already worked out in [6]: 1. Eciency: P p j=1 ˚ j= v GE(P). All SFIMP values add up to v GE(P), i.e., the dierence in perfo”
Data Shapley: Equitable Valuation of Data for Machine Learning (2019, background): “Shapley value has been recently proposed as a feature importance score for the purpose of interpreting black-box predictive models (Kononenko et al., 2010; Datta et al., 2016; Lundberg & Lee, 2017; Cohen et al., 2007; Chen et al., 2018; Lundberg et al., 2018).”
An Efficient Explanation of Individual Classifications using Game Theory (2010, methodology): “Their MSA framework is later used and adapted into a method for feature selection (Cohen et al., 2007).”
Strengthening learning algorithms by feature discovery (2012, methodology): “Feature selection is used in various new machine learning classifiers, including learning to select features using their properties [11], feature selection for unlabeled data [12], selection using ensemble methods [12], feature combination by multimodal learning [26], genetic feature selection [3], ”
Shapley values for feature selection: The good, the bad, and the axioms (2021, unknown): “Alternatives to Algorithm 1 are found in [40]–[43].”