New Clustering Algorithm for Multidimensional Data

Calculating similarity for multidimensional data is one of the key problems that must be addressed in order to promote the development of data clustering algorithms. In this study, we developed and tested a new similarity calculation index to improve the accuracy of multidimensional data clustering. First, the information divergence (ID) and generalized gradient angle (GGZ) were explored in detail. Second, the ID and GGZ were combined to calculate the similarity of multidimensional data, thus enabling a new algorithm for data clustering. Finally, two experiments were conducted to evaluate the performance of the proposed algorithm. The results of the experiments demonstrate that our proposed similarity calculation index for multidimensional data is both accurate and effective, providing better performance as measured by the metrics of accuracy (ACC), normalized mutual information (NMI), and purity (PUR). Based on this research, we conclude that the application of the proposed similarity calculation index is conducive to the improvement of data clustering for multidimensional data. Subject Categories and Descriptors K.2.8 [Database Applications]: Data clustering; B.2.4 [HighSpeed Arithmetic]: Algorithms General Terms Data clustering, Algorithm