Research of Artificial Neural Network Algorithm Based on Fuzzy Clustering of Rough Sets
BPNN (back propagation neural network) has advantages of good learning and memory, but it hasnt the ability to train samples with qualitative attributes. On the contrary,fuzzy clustering of rough sets based on fuzzy equivalent matrix has the advantage. In the paper, Fuzzy clustering of rough sets is used as anterior processor of BPNN,which is to say that firstly data records are partitioned by fuzzy clustering of rough sets and then are inputted into BPNN to be learned and memorized. BPNN model based on fuzzy clustering of rough sets and model of Multivariate Linearity Regression are both applied in costume field. By comparing the results of BPNN model with that of Multivariate Linearity Regression model, the results show that the BPNN model based on fuzzy clustering of rough sets is more effective and practical.
Artificial neural network BPNN Fuzzy clustering Fuzzy Equivalent Matrix Fuzzy Similar Matrix
Yan Chen Taoying Li
School of Economics and Management, Dalian Maritime University Dalian, Liaoning, China
国际会议
杭州
英文
744-747
2006-10-12（万方平台首次上网日期，不代表论文的发表时间）