Chaotic BP Neural Network Based Case Retrieval Method
Chunguang Chang, Dingwei Wang, Yachen Liu · 2006
The principle of chaotic BP neural network is described, the control strategy of annealing strategy is introduced so as to control the chaotic movement during the later training, and some related conclusions are proved. Then, the implement steps of chaotic BP neural network based case rerieval (CR/CBPNN) method are introduced. An improved Delphi approach based algorithm is proposed to determine output sample set so as to improved the quality of training. To validate the efficiency of CR/CBPNN method, the time complexity is calculated. The simulation result shows that compared with the nearest neighbor (NN) algorithm it can improve case retrieval precision for complex system such as dynamic scheduling system for steel-making.