Fusion of Accident Recurrence Results by Improved Fuzzy Entropy Method Based on Clustering
On the basis of uncertainty analysis,a fusion method of multi-model results based on fuzzy trans-formation is proposed in this paper.Firstly,the results of multi-model accident reproduction are discretized,the clustering results are obtained by K-means clustering,and the extreme values of each cluster interval are used as sub-intervals.According to the entropy method,the reliability of each model is obtained.The minimum informa-tion entropy principle is used to integrate the D-S evidence theory into the model reliability of each expert.Final-ly,the reliability of the model is used as the weight set and the evaluation matrix to make a comprehensive evalu-ation of fuzzy transformation,and the final fusion result is obtained.And through a real accident case to carry on the practical application.
Accident reconstructionEntropy methodD-S evidence theoryMinimum information entropyFuzzy mapping
万方数据
维普
