Fuzzy pattern recognition method for assessing soil erosion.

In this paper a PSIAC-based multi-parameter fuzzy pattern recognition (MPFPR) model is proposed and applied for classifying and ranking the potential soil erosion (PSE). In this approach, standard value matrix is used to define the membership degrees of each catchment to each class and the feature values are used for alternative ranking. The characteristic of PSE for each class is expressed by linguistic variables. The proposed method is straightforward, easy to understand, very practical, and its results may easily be interpreted. To assess the performance of the model, the results of PSIAC MPFPR and original PSIAC method are interpreted and compared with the observed data. It is shown that the proposed approach reflects the fuzzy nature of the soil erosion more efficiently and is quite robust for application in real world cases.