A fuzzy pattern recognition model for water quality evaluation based on the principle of maximum entropy

摘要

A fuzzy pattern recognition model for water quality evaluation was developed on the basis of the least weighted general distance and the objective influence of uncertainty estimated by the principle of maximum information entropy of Jaynes, To balance the least weighted general distance and the maximum entropy, a new methodology was developed to use the model of maximum entropy fuzzy pattern recognition as the practical modelling process, to use the grade judgment standard as the principle of theoretic grade, and to use accelerating genetic algorithms to determine the balanced parameter a between the least weighted general distance and the maximum entropy. The theoretical analysis and applications show that the new method for determining the balance parameter a is highly feasible and reliable. The new model is theoretically sound and widely applicable for fuzzy pattern recognition for handling various problems in comprehensive water resources evaluation.