Machine Learning in Manufacturing: Processes Classification Using Support Vector Machine and Horse Optimization Algorithm

Abstract

The classification of the manufacturing processes in processes that pass the in-line testing and processes that fail the in-line testing is a challenging research problem as the manufacturing processes data is characterized by many features that correspond to the different steps of the manufacturing processes. This research article proposes a method in which: (1) the manufacturing processes classification is performed using the Support Vector Machine (SVM) algorithm, (2) the regularization parameter value and the gamma coefficient value of the SVM algorithm are optimized using Horse Optimization Algorithm (HOA), (3) the HOA based SVM results are compared to Particle Swarm Optimization (PSO) based SVM results and Chicken Swarm Optimization (CSO) based SVM results, and (4) the data used in experiments is the open source public dataset SECOM.