I-WOA: An Optimization of K-Means Clusteringsis

As the big data technology is widely used in various research fields, cluster analysis has been generally regarded as one of the most effective methods for processing different types of data. In response to the shortcomings of the K-means algorithm, such as sensitivity to the initial clustering points and a propensity to fall into local optima, an Improved Whale Optimization Algorithm (I-WOA) is proposed to optimize the K-means clustering algorithm. In the proposed I-WOA algorithm, the Sobol sequence is introduced to initialize the whale population, thus enriching the diversity of the whale population. An elite reverse learning strategy is also employed to improve the position updating method of the optimal whale in the whale population, hence enhancing the quality of the solution. Additionally, the convergence factor(\(a\)) and contraction probability (\(p\)) in the whale algorithm are adjusted to balance global and local searches, therefore improving convergence speed and optimization precision. Finally, the I-WOA algorithm is evaluated by six famous benchmark functions. The experimental results show that the proposed algorithm has better cluster quality and global search capability compared to traditional clustering algorithms, and outperforms some popular algorithms.