An Extended Grid-based Clustering Algorithm with Referential Value of Parameters

摘要

GRPC algorithm (a grid-based clustering algorithm with referential parameters) had put forward by authors of this paper, and the algorithm provided user with a feasible technology for reducing the blindness of parameter assignment. The work of this paper is to develop GRPC algorithm in the aspects of scalability and high-dimensionality. We process large-scale data set by means of random sampling technique and transform the clustering of high-dimensional data into the clustering of two-dimensional data. Experimental results confirmed that this algorithm (EGRPC) could effectively cluster vast data set whose data is three-dimensional or other high-dimensional.