Firstly, the spatial distribution graphic of two-dimensional membership function for 4 types of standard site was presented. It can be used in fuzzying the site classification directly. The site classifications and their fuzzy characteristics are given much consideration. Then, a continuity method of site classification methods based on the two-dimensional membership function was provided. Nextly, by using spatial distribution graphic of continuous classification above, the method of site continuous classification based on neural network (NN) is established. Finally, the applications are given. The practical results show that this method realized site intelligent continuous classification. This method gets rid of the illogicality of discontinuity and appreciable distinction displayed on the code chart between the site edges, as well as overcomes disadvantage of intelligence level and lower efficiency, those exit in traditional classification. This method causes the spatial distribution of site classification to be more smoothly, and is more corresponding to the practical engineering.
[1]
Albert Y. Zomaya,et al.
Adaptive Model-Based Control using Neural Networks
,
1993
.
[2]
Jinping Ou,et al.
Intelligent Continuity Method of Site Classification and Characteristic Period
,
2006
.
[3]
Wang Guang-yuan.
Site classification method and its application based on two dimensional membership function
,
2005
.
[4]
Alan F. Murray,et al.
IEEE International Conference on Neural Networks
,
1997
.
[5]
Sibel Pamukcu,et al.
Site and Geomaterial Characterization
,
2006
.
[6]
R. Hecht-Nielsen,et al.
Theory of the Back Propagation Neural Network
,
1989
.