MULTILAYER PERCEPTRON CLASSIFICATION MODEL BASED ON WEIGHTED MAJORITY VOTING FOR CLASSIFYING ANCIENT GLASS PRODUCTS

Abstract

This study explores the application of weighted majority voting combined with machine learning and artificial neural networks to the categorization of ancient glass artifacts. The study developed an optimized Multilayer Perceptron (MLP) classification model based on the cyclically optimized results from various well-performing classification models through weighted majority voting. The model demonstrated strong performance in cross-validation tests, achieving a prediction accuracy of 92.75%, demonstrating high stability and precision. This research provides a novel approach and methodology for the classification of ancient glass artifacts, potentially contributing to further advancement in this field