Grain Quality Evaluation Method Based on Combination of FNN Neural Networks with D-S Evidence Theory

The output of the fuzzy neural network was adopted as BPAF (basic probability assignment function) in this paper. By training the fuzzy neural network, the massive language fuzzy information and the concerned expert's experience were integrated in the decision process, it advantageous in enhancing the BPAF of the accuracy, the reliability and the objectivity. Therefore, using the superiority of D-S evidence theory in the processing uncertainty aspect and analyzing the situation of grain by FNN and the D-S evidence theory union, it can greatly decrease the uncertainty of system.