A method of dynamically determining the number of clusters and cluster centers

Text clustering is an important technology in the field of data mining. The traditional K-means algorithm is sensitive to the number of clusters, and there is a limitation that the result of randomly initializing cluster centers is not stable. This paper presents a method of dynamically determining the number of clusters and cluster centers based on text similarity matrix. The experiment results show that the method works well and improves the K-means algorithm's accuracy and adaptability.