The perform of k-means based on hybrid artificial bee colony and genetic algo rithm in optimizing the cluster size of k-means

This study tackles the growing complexity of data by presenting a novel approach to KMeans clustering: the Artificial Bee Colony (ABC) and Genetic Algorithms (GA) KMeans algorithm. The traditional K-Means method has inherent weaknesses, such as arbitrary cluster selection and random initialization of cluster centers. This research addresses these issues by focusing on determining the optimal number of clusters in unlabeled data, a key requirement for effective clustering. The proposed ABC_GA_KMeans algorithm overcomes these challenges through autonomous optimization of data collection and cluster centers. It achieves this by integrating ABC optimization and GA to solve the binary optimization problem often encountered in K-Means clustering. Additionally, the inclusion of Genetic Neighbourhood Generators (GNG) enhances the algorithm's ability to compare results within the ABC network, contributing to improved robustness and efficiency. The study conducts extensive experiments with both simulated and real-world datasets to evaluate the performance of the ABC_GA_ K-Means algorithm against conventional clustering techniques, including traditional KMeans, Fuzzy K-Means, and other approaches. The results demonstrate that the proposed algorithm consistently outperforms these methods in terms of accuracy, precision, recall, and rand index. Notably, the ABC_GA_K-Means algorithm achieved high accuracy on datasets like Zoo: 0.9119, Breast Cancer: 0.9413, and Soybean: 0.9575, underscoring its effectiveness in optimizing data collection and cluster centers. These results not only validate the robustness of the proposed algorithm but also highlight its versatility, making it suitable for a wide range of applications. Implication of this innovative approach to clustering points to the potential for further research into alternative heuristics within the ABC_GA_K-Means framework, offering new avenues for advancing the field of clustering. Keyword: Artificial Bee Colony, Clustering, Genetic Algorithm, Gendata, Initial number of Cluster, K-Means