Kernel-based Information Criterion

TL;DR

This paper presents Kernel-based Information Criterion (KIC), a new model selection method for regression that uses kernel-based complexity to improve robustness and performance over existing techniques.

Contribution

Introduction of KIC, a kernel-based complexity measure for model selection that enhances robustness and outperforms existing criteria in regression tasks.

Findings

KIC outperforms LOOCV, ICOMP, and GPR in experiments.

KIC effectively measures parameter interdependency using variable-wise variance.

Experimental results demonstrate superior performance on simulated and real data.

Abstract

This paper introduces Kernel-based Information Criterion (KIC) for model selection in regression analysis. The novel kernel-based complexity measure in KIC efficiently computes the interdependency between parameters of the model using a variable-wise variance and yields selection of better, more robust regressors. Experimental results show superior performance on both simulated and real data sets compared to Leave-One-Out Cross-Validation (LOOCV), kernel-based Information Complexity (ICOMP), and maximum log of marginal likelihood in Gaussian Process Regression (GPR).