Research On K-means Method Based On Cuckoo Algorithm

As Big Data era emerges and technologies related to cloud computing gradually improves, data witnesses an explosive growth. Big data has such traits as large volume, diversification, large value and high frequency, etc. Meanwhile, people think more highly of data, and they expect to discover useful and valuable information from massive data. As a kind of data mining algorithm, clustering analysis has gained much attention. It has been successfully applied to many fields, like data mining, Pattern recognition, computer vision, image segmentation, information retrieve, E-commerce, biological industry, insurance industry, medical detection.Maintenance of process model library is a hot topic of business process management, and process model clustering is one method to address the problem. Process model clustering not only helps corporation leaders make decisions, but also helps users customize their personalized process, which improves the use efficiency and user satisfaction of process model.As a classical partition-based clustering method, k-means is easy to implement, so it is widely used in the field of data mining. With the development of data mining techniques and the emergence of novel swarm intelligence optimization algorithms, more and more scholars focus on combining k-means with these optimization algorithms. K-means algorithm has some typical drawbacks, like dependence on initial solution, easy to be strapped into local optimal solution, etc. To address these drawbacks and improve the traditional K-means algorithm, this dissertation combines cuckoo algorithm with K-means. Simulation experiment results show that the new method has relatively high feasibility and effectiveness.