A modified fuzzy C-Means algorithm based on gravity and cluster merging

作者：Zhong Jiang^*; Liu Longhai; Chen Qiang; Chen Xue; Zhou Ying

来源：Journal of Information and Computational Science, 2010, 7(13): 2699-2706.

• Cluster Merger
• Cluster-merging
• Clustering techniques
• Experimental evaluation
• FCM Algorithm
• Fuzzy C means clustering
• Fuzzy C-means algorithms
• Fuzzy Cluster
• Gravity
• Initial cluster centers
• Modeling results
• Optimum number

There exists a large number of clustering algorithms in data mining. Fuzzy C-Means (FCM) clustering is one of the most well-known and commonly used clustering techniques. It has the advantage of producing good modeling results in many cases. However, it can not specify the optimum number of clusters by itself. In addition, it is sensitive to the initial cluster centers and outliers. To overcome the above drawbacks, we propose a modified FCM Algorithm based on gravity and cluster merging in this paper. In this algorithm, the initial cluster centers were selected and the influence of outliers was minimized by using gravity, and an optimum number of clusters could be specified by using cluster merging. The experimental evaluation shows that the modified method can effectively improve the clustering performance.

• 出版日期2010