Abstract:In order to explore the influence of the angle between elements on the incremental clustering algorithm

this paper is inspired by angle based outlier detection algorithms. The paper establishes a dynamic incremental clustering algorithm model based on angle measurement

and uses the angle between elements to reasonably distribute the increased data. For the KFCM algorithm to randomly obtain the defects of the cluster center

the paper uses the average distance to improve the algorithm so that the KFCM algorithm can automatically obtain the cluster center. Since the kernel function has a great influence on the clustering result

further studies have been done on how to select the kernel function. The experimental results show that this algorithm can effectively process various data sets for increasing data.