THE FEATURE WEIGHTED FCM ALGORITHM WITH SEMI-SUPERVISION
The traditional FCM algorithm is an unsupervised fuzzy clustering, but in real life, there are a lot of known knowledge, and a large number of samples have known information, how to take full advantage of these known information of cluster become a hotspot of research. If the known knowledge is added to the FCM algorithm in the optimal problem, we will get an optimal problem with some constraint conditions. Throng appropriate variable substitution and the thought of HPR (Hestenes-Powell-Rockafellar) algorithm, we obtain the feature weighting FCM algorithm with semi-supervision. Because the algorithm is compared with the original FCM algorithm, the number of variables is not increased, thus there is little effect on its speed of operation. IRIS data experiment shows that the algorithm not only deepens the scope of discussion of the semi-supervised FCM algorithm, but also makes the computation complexity little. Compared with the existing semi-supervised FCM algorithm, the new algorithm has greater improvement, and provides a way of thinking for the FCM algorithm with supervision, we will discuss in another text.
FCM The feature weighted Semi-supervision
TONG Xiao-Jun JIANG Qin SANG Nong GAN Hai-Tao ZENG Shan
Wuhan Polytechnic University Wuhan,Hubei,430074, China
国际会议
武汉
英文
22-26
2009-10-16（万方平台首次上网日期，不代表论文的发表时间）