Overview

Feature selection studies which variables, channels, bands, or learned dimensions carry useful information, aiming to simplify models and improve prediction, interpretation, or measurement efficiency. Common approaches include Lasso and elastic net regularization, filters such as mRMR and Boruta, Bayesian spike-and-slab models, and wrapper or evolutionary searches; newer methods use neural gates, attribution scores, and sequential policies. Unsupervised methods select features for clustering or representation learning, while applications span tabular data, omics, EEG, hyperspectral imagery, and time series. A central challenge is that selected features can change with correlated inputs, evaluation design, or distribution shifts, so predictive accuracy must be considered alongside stability, error control, computational cost, and strong baselines.

Latest papers

CardsList

213 papers23 papers in the last four weeks, down 12% on the four weeks before. 0.2% of all new papers.

  1. Feature Selection for Latent Factor Models

    Dec 13, 2024Rittwika Kansabanik, Adrian BarbuFeature Selection

  2. Heterogeneous Graph Attention Network With Multimodal Feature Selection for Cancer Multiomics Integration

    Aug 5, 2024Sina Tabakhi, Charlotte Vandermeulen, Ian Sudbery +1Multi-Omics Data IntegrationHeterogeneous GNNs

  3. Regularized Estimation and Feature Selection in Mixtures of Generalized Linear Experts

    Jul 14, 2019Thin Nguyen-Van, Faicel Chamroukhi, Ha Hoang Van +1Mixture of ExpertsSparse Mixture-of-Experts