A Density-based Improvement for Distributed Clustering Algorithm

The density based distributed clustering algorithm (DBDC) has been employed widely in process of distributed clustering.A major drawback of DBDC is the high time complexity.This work presented a method based on representative point for global clustering.A data grid mapping method which mapped data object to space grid, which improved DBDC and enhance the space search process and efficiency of implementation in local clustering level.Clustering error can be reduced by aglobal clustering method based on representative point.Using verification experiment with real data set from open source software, results showed this algorithm can improve original DBDC in two aspects: accuracy and reducing time complexity.