An improved seedless clustering algorithm based on the average correlation

作者：Cui Licheng^*; Zhang Weishi; Zhai Huawei; Zhang Runtong; Zhang Xiuguo

来源：Journal of Information and Computational Science, 2012, 9(8): 2287-2294.

• Clustering number
• Correlation cluster
• Correlation clustering
• Correlation coefficient
• ISACC

Clustering algorithm can group similar objects, but it is a little difficult to find incidence relation between objects. So, finding correlation clusters is important in some fields and most of current correlation clustering algorithms are sensitive to the initial set of seeds or initial clustering number. Based on these, an improved seedless clustering algorithm based on the average correlation (ISACC) is proposed, which is referring to the idea of ROSCC algorithm. ISACC is based on the PCA and average correlation coefficient, it could avoid the effects of initial set of seeds or initial clustering number, and also handle exceptional information. Experimental results on both synthetic demonstrate its effectiveness and accuracy.

• 出版日期2012