Application of Fuzzy Pattern Recognition in Prediction of Karst Collapse Hazard

摘要

It was known that the condition effecting karst collapse was very complicated. Influential factors were various and interdependent, geologic body itself was heterogeneous and uncertain, and human activities were changeful. Thus, it had great application value and theoretical significance that introduced the fuzzy theory in prediction of karst collapse hazard. Fuzzy pattern recognition was used to predict the covered karst collapse hazard in Wuhan. Since existing closeness degree couldn't recognized efficiently, weight value was mingled with closeness degree and generalized Euclidean weight distance closeness degree was defined. In order to research, thirteen main factors of influencing karst collapse's hazard were selected as the evaluation index according to the index table of predicting karst ground collapse (Chen, 1994). Furthermore, weight value of each index could be obtained by means of analytic hierarchy process (AHP). By the method, karst collapse risks of 1631 cells in the study area were predicted. Moreover, the result of recognition was in concordance with the fact. Accordingly, it verified that the model of fuzzy pattern recognition with generalized Euclidean weight distance closeness degree was very feasible to predict the covered karst collapse hazard in Wuhan.