模糊逼近算法与人工神经网络预测功能

有些实际问题是无法用已知的很多定量预测方法预测的，这更是在预测中有极大局限性的人工神经网络难以解决的问题．本文通过对预测问题模糊逼近算法的研究，提出新的模糊逼近泛函微分方程定量预测方法，并从而化为能用人工神经网络预测的方法，扩充人工神经网络解决实际问题的功能．

Fuzzy Proximation Algorithms and Practical Forecasting Function by Artificial Neutral Networks

There are some practical forecasting problems which have no way of forecasting by many known measurable forecasting methods, and these problems are more difficulty to be solved by the Artificiol Neutral Network, which is more limited in practical forecast. By studying fuzzy proximation algorithms for forecasting problems, a new measurable forecasting method of the functional differential equations in term of fuzzy proximation has been proposed and thereby can be turned into forecastable method in terms of Artificial Neutral Network which extends the function of solving practical problems by Artificial Neutral Network.