Towards an automated method based on Iterated Local Search optimization for tuning the parameters of Support Vector Machines

TL;DR

This paper introduces a heuristic optimization method based on Iterated Local Search to automatically tune Support Vector Machine parameters across different datasets, aiming to improve model performance.

Contribution

It presents a novel heuristic approach using Iterated Local Search for automatic SVM parameter tuning, enhancing adaptability to new datasets.

Findings

Preliminary formulation of the optimization strategy.

Potential for improved SVM parameter tuning.

Framework adaptable to various datasets.

Abstract

We provide preliminary details and formulation of an optimization strategy under current development that is able to automatically tune the parameters of a Support Vector Machine over new datasets. The optimization strategy is a heuristic based on Iterated Local Search, a modification of classic hill climbing which iterates calls to a local search routine.