Optimization of Support Vector Machine using Firework Algorithm for Breast Cancer Classification

The goal of this work is to enhance the realization and performance of Support Vector Machine (SVM) models by optimizing their hyperparameters using the Fireworks Algorithm (FWA). This paper addresses testing data from the Breast Cancer Wisconsin (Diagnostic) dataset compared with classical SVM models using models optimized with FWA. The FWA-optimized SVM gets much higher classification accuracy than the basic SVM, with the adjustment of the basic parameters optimum for the regularization constant, C, and the parameters of the kernel. Such results indicate the applicability of FWA in enhancing the performance of SVM and unveil its potential for hyperparameter optimization in different fields.