Bilevel hyperparameter optimization for support vector classification: theoretical analysis and a solution method

TL;DR

This paper reformulates hyperparameter tuning for support vector classification as a bilevel optimization problem, converting it into an MPEC and proposing a global relaxation algorithm with proven convergence, demonstrating superior performance in experiments.

Contribution

It introduces a novel bilevel optimization framework for hyperparameter selection in SVC and develops a convergent global relaxation algorithm tailored for this problem.

Findings

The proposed algorithm converges to a C-stationary point.

The MPEC satisfies the MFCQ condition at feasible points.

Numerical results show superior generalization performance.

Abstract

Support vector classification (SVC) is a classical and well-performed learning method for classification problems. A regularization parameter, which significantly affects the classification performance, has to be chosen and this is usually done by the cross-validation procedure. In this paper, we reformulate the hyperparameter selection problem for support vector classification as a bilevel optimization problem in which the upper-level problem minimizes the average number of misclassified data points over all the cross-validation folds, and the lower-level problems are the l1-loss SVC problems, with each one for each fold in T-fold cross-validation. The resulting bilevel optimization model is then converted to a mathematical program with equilibrium constraints (MPEC). To solve this MPEC, we propose a global relaxation cross-validation algorithm (GR-CV) based on the well-know Sholtes-type global relaxation method (GRM). It is proven to converge to a C-stationary point. Moreover, we prove that the MPEC-tailored version of the Mangasarian-Fromovitz constraint qualification (MFCQ), which is a key property to guarantee the convergence of the GRM, automatically holds at each feasible point of this MPEC. Extensive numerical results verify the efficiency of the proposed approach. In particular, compared with other methods, our algorithm enjoys superior generalization performance over almost all the data sets used in this paper.