From Linear Programming Approach to Metaheuristic Approach: Scaling Techniques

Abstract

The objective of this work is to propose ten efficient scaling techniques for the Wisconsin Diagnosis Breast Cancer (WDBC) dataset using the support vector machine (SVM). These scaling techniques are efficient for the linear programming approach. SVM with proposed scaling techniques was applied on the WDBC dataset. The scaling techniques are, namely, arithmetic mean, de Buchet for three cases $\left(p=\mathrm{1,2},\text{and\hspace{0.17em}}\infty \right)$ , equilibration, geometric mean, IBM MPSX, and Lp-norm for three cases $\left(p=\mathrm{1,2},\text{and\hspace{0.17em}}\infty \right)$ . The experimental results show that the equilibration scaling technique overcomes the benchmark normalization scaling technique used in many commercial solvers. Finally, the experimental results also show the effectiveness of the grid search technique which gets the optimal parameters (C and gamma) for the SVM classifier.