|
|
|
| Improved RBF Regression Using Fuzzy Partition and Supervised Fuzzy Clustering |
| Chen Cong; Wang Shi-Tong |
| School of Information Technology, Jiangnan University, Wuxi 214122, China |
|
|
|
|
Abstract In order to improve the precision of RBF regression, this article advances a novel RBF regression modeling method using fuzzy partition and supervised clustering. The proposed method first splits the training data into several subsets using supervised clustering. Then local regression models are independently built with RBF network for each subset. Finally, the output of the network is formed with a weighted combination of each local model. Experiments show that the proposed method achieves more accurate interpretation of local behavior of the target model.
|
|
Received: 31 March 2008
|
|
|
|
|
|
|
|
|
2009, Vol. 31 
