Adaptive K-valued K-means clustering algorithm
Wang Sheng-hui, Li Hanbing · 2020 5th International Conference on Mechanical, Control and Computer Engineering (ICMCCE) · 2020
Based on particle swarm optimization (PSO), the algorithm for selecting appropriate K values is improved by combining k-mean algorithm. When the algorithm converges, the expansion and reduction of K value can be determined by comparing the relationship between different K value selection and global optimal fitness. Experiments show that the improved algorithm can assist K value selection effectively and obtain a better clustering effect.