科研之友
微信
新浪微博
Facebook
分享链接
摘要
In order to solve the practical issues of emergency material undersupply and waste in the management of emergency effectively and realize the valid prediction for demand of emergency supplies, the Index Screening based on Rough Set Theory is presented. Such method can result in a reduced indication set by Domain Partition and Attribute Reduction of the fuzzy similarity relation of original index. In order to improve the accuracy of prediction results, establishing the multi input multi output ( MIMO) prediction model of wavelet network and determining the variable parameter of prediction network by adopting method of iterative steepest descend and gradually test play better, the characteristics of wavelet network such as self-learning, local approximation and adaptive resolution and fault-tolerant etc. are studied. Finally, by a series of historical seismic data as the training sample, the structure of prediction model is determined. By contrasting the prediction results and the actual results, the effectiveness of the model is verified.
- 出版日期2013-6
- 单位重庆大学
