The Improvement and Implementation of dBscan Clustering Algorithm
Yang Liu · Microelectronics & Computer · 2009
Firstly,a description and an analysis of dBscan clustering algorithm based on density were presented,a methodology was also described to determine the value of parameter and parameter MinPts,the features of algorithm dBscan were then discussed and meanwhile the defect of dBscan was pointed; and then,dBscan was improved from two aspects:one is from the process of non-spatial attributes,and the other is from the acceleration of clustering speed,as a result,the search tree generation process was simplified and the efficiency of dBscan algorithm was improved; at last,a framework of implementation of the improved dBscan algorithm was described in detail.