A New Hybrid Clustering Algorithm Based on Stimulated Annealing
Chengji Zha, Yinan Dou, Minjie Guo, Yue Dong · 2013
In the recent years, more and more researches are preferred to focus on network user behavior. Usually, k-means clustering and Agglomerative Nesting (AGNES) are respectively chosen to analyze the network user behavior. But both the two kinds of algorithm have some disadvantages inherently. A kind of hybrid clustering algorithm (ASAKM) is proposed in this paper, which takes the advantages of both kinds of clustering algorithms. Furthermore, the idea of simulated annealing is also adopted in this paper, to implement the global optimal solution while the partitioning methods usually only reach the local optimal minimum. Experiments indicate that, with this new hybrid algorithm, the clustering results can be more accurate.