Feature selection with scalable variational gaussian process via sensitivity analysis based on L2 divergence

Published: 01 Jan 2023, Last Modified: 15 May 2025Neurocomputing 2023EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•A new model agnostic feature selection method for supervised learning.•Sensitivity analysis with the predictive distribution of a scalable Gaussian process.•New L2 divergence derivation to better measure classification results and changes.•Better performance on several learning models than existing methods.
Loading