模糊综合评价的一种神经网络方法

将神经网络理论应用于模糊综合评价，首先引入因素空间的概念，在因素空间中给出了一种基于三层神经元的多用量化模糊综合评价方法，并利用LMS算法及矩方法研究了该模型权重及学习步长的确定方法，实现了权重及学习步长的网络学习，最终实现了等级评价值的计算机输出．较好地解决了传统的综合评价方法中各指标的权数受人为因素的影响这一难点问题．

A Method of Neural Network on Fuzzy Multifactorial Evaluation

The theory of neural network was applied to fuzzy multifactorial evaluation. We introduce the concept of factor spaces, and proposed a method of fuzzy multifactorial evaluation with multi-quantification on three-layer neuron. We also discussed how to make decision on weights and learning-step using the LMS method and the moment method, and implemented the output about the value of evaluation degree. In the traditional methods of multifactorial evaluation, the multi-indexes weighters may be affected by man-made factors, but this method overcomes to some extent this problem.