摘要
模糊C均值聚类具有较广泛的应用,但该聚类算法本身存在容易陷入局部最优、对初始值敏感的缺点.本文提出基于蝙蝠算法与模糊C均值算法相结合的BAFCM聚类算法,并通过数值实验对比,说明BAFCM聚类效果优于FCM、PFA.
Fuzzy C-mean clustering is widely used, but the FCM is easy to fall into local optimum shortcomings, sensitivity to the initial data. The BAFCM method is proposed,which based on the combining of the BA and FCM algorithm. Finally through the contrast of FCM, PFA, BAFCM clustering algo- rithm, it is found that the new algorithm is better than the FCM and PFA.
出处
《西安工程大学学报》
CAS
2013年第5期680-683,693,共5页
Journal of Xi’an Polytechnic University
基金
陕西省软科学基金项目(2012KRM58)
陕西省教育厅自然科学基金项目(12JK0744
11JK0188)
