A Hybrid Framework for Ancient Glass Classification and Recognition: Combining Binary Nutcracker Optimization Algorithm with KNN Method

Accurately categorizing ancient glass artifacts, which undergo weathering in burial environments and experience compositional changes, is a challenging problem. This paper presents a novel approach that combines the Binary Nutcracker Optimization Algorithm (BNOA) with the k-Nearest Neighbor (KNN) method for the classification and recognition of ancient glass artifacts based on chemical composition analysis. The approach aims to accurately categorize ancient glass by leveraging BNOA for dimension reduction and feature extraction, while KNN is utilized for classification and recognition tasks. The proposed hybrid method is evaluated on 10 standard benchmark datasets from the UCI repository and a real dataset containing 67 samples of Ancient Chinese glassware using a 5-fold cross-validation approach. The results demonstrate that the BNOA-KNN model enhances the classification performance of KNN and achieves higher accuracy compared to other methods such as SVM, NB, ANN, and RF. Furthermore, the proposed BNOA-KNN approach is applied to predict the categories of eight unknown ancient glass samples, showcasing its practical applicability in real-world scenarios.