Efficient Grid-based Clustering Algorithm with Leaping Search and Merge Neighbors Method

The increasing data size makes the research of clustering algorithm still an important topic in data mining. As one of the fastest algorithms, the grid clustering algorithm now still suffers from low precision problem. And the efficiency of the algorithm also needed improvement. In order to cope with these problems, this paper proposes an efficient grid-based clustering algorithm by using leaping search and Merge Neighborhood (LSMN). In the algorithm, the LSMN first divides the data space into a finite number of grids and determines the validity of the grid according to the threshold. Then, leaping search mechanism is used to find valid grids of the grid by retrieving all the odd columns and odd rows. Finally, if the number of valid grids is greater than the invalid grid, the invalid grids are merged together. In the algorithm, the time cost is reduced and the accuracy is improved by leaping search and re-judgment of the invalid grid mechanisms respectively. The experimental results have shown that the proposed algorithm exhibits relatively better performance when compared with some popularly used algorithms.