Research And Improvement On Density Peak Clustering Algorithm And Application For Earthquake Classification

The science and technology are the vital factor for the transformation of the human society from manufacture to information.Nowadays the information have become the crucial resource for social production.And how to process the big data with terabytes has become the most notable problem in data mining field.As a popular and useful method in data mining field,the clustering has attracted more and more researchers all over the world.The Density Peak Clustering has been published by Rodriguez and Laio in the Science since 2014.The DPC has been recognized by various field widely,and it is a outstanding algorithm for finding cluster centers fast and effectively.But there still are many disadvantages exist in the DPC:(1)The DPC is unable to deal with the data points with low density in the data set.The algorithm will distribute the anomalies or hub nodes to the clusters mistakenly.(2)The process of selecting cluster center is determined by human that may decrease the objectivity and accuracy of the clustering result.(3)The algorithm can hardly gain a good clustering when it meets the complex structure data such as flow pattern,different density and various scale data.In view of the above problems we put forward some methods to improve the original DPC:(1)In order to overcome the defect of the original DPC in detecting anomalies and hub nodes,we explore its potential reasons of adopting halos instead of low density nodes,and proposed an improved recognition method on halo node for density peak clustering algorithm(HaloDPC).The proposed HaloDPC serves as an excellent enhancement of DPC to improve its capability with varying densities,irregular shapes,cluster number,outliers and hub nodes detection.We demonstrate the power of the HaloDPC algorithm on several test cases.(2)In view of the low objectivity and accuracy because of the man-made factor,we proposed a density fragment clustering without peaks.This proposed algorithm was inspired from DPC,DBSCAN and SCAN to satisfy more synthetic data set.To put the local density in descending order and build the cluster fragment,this method can effectively find the true cluster center automatically.Then the data points will be classified by the structural similarity finally.(3)In view of the unsatisfying clustering effect of density peak clustering algorithm when dealing with data sets of complex structures,a semi-supervised affinity propagation clustering algorithm based on density peaks(SAP-DP)was proposed in this paper.This algorithm combined the advantage of the density peak and affinity propagation,then use the semi-supervised pairwise constraint information to adjust the similarity matrix.(4)We apply this method to earthquake classification,the simulation results shows that the proposed algorithms can classify the earthquake level correctly and it will has a big potential in actual application.Besides we excavate the disadvantage of the proposed algorithms in depth that will help to improve the algorithms more valuable and feasible.