Research On Hybrid Algorithm Based On Subtractive Clustering

Due to the database technology has become increasingly perfect has brought a sharp increase in the amount of data, people have accumulated quite a lot of data, in a large amount of information environment, people feel difficult to find for yourself useful knowledge and information. How to dig out the useful information in the fast-growing mass of data is a necessary requirement of many scholars present research objectives, as well as the information technology development.Cluster analysis of data mining technology is an important branch, it can handle data alone, but also used in combination with other algorithms that can effectively understand the distribution of data, looking for some hidden features in the data. In this paper, some clustering algorithm deficiency exists, combined with subtractive clustering method is proposed to improve the performance of clustering algorithms, specific research includes the following aspects:(1) This paper describes the clustering basic knowledge and concepts, several research directions of cluster analysis, the similarity measure function of clustering and the criteria function of clustering to assess the clustering quality, studied the classification clustering and the most classical clustering algorithm features and procedures, finally comparative analysis the performance of various types of clustering algorithms, the advantages and disadvantages of various clustering algorithms and analysis the algorithm suitable range.(2) This paper presents an mixed subtraction clustering improved hierarchical clustering algorithm. First,using subtractive clustering to get the initial clustering results, and then use the minimum spanning tree Kruskal algorithm to find the optimal path, according to the weight of the distance between the classes, and finally achieve the hierarchical clustering. Through simulation experiments of the UCI datasets, verify the effectiveness and superiority of the proposed algorithm. Experimental results show that the improved hierarchical clustering based on subtractive clustering algorithm to perform fast, high efficiency, and the clustering effect is better than traditional clustering algorithms. In particular, the more the amount of data, more obvious this algorithm advantage in terms of time consumed.(3) This paper proposes an mixed subtraction clustering improved affinity propagation clustering algorithm. First, add the subtraction clustering, using the density value of the data points to obtain the point of initial clusters. Then, calculate the similarity distance between the initial cluster points, and reference the idea of semi-supervised clustering, adding pairs restriction information, structure sparse similarity matrix. Finally, the cluster representative points conduct AP clustering until a suitable cluster division. Experimental results show that the algorithm reduces the amount of storage similarity matrix, reducing the amount of computation, and better than the original algorithm on the clustering effect and processing speed.