基于神经网络的专家预测方法研究

A method research of the experts prediction based on artificial neural network

针对复杂预测问题的需求,基于神经网络的思想对专家预测方法进行了探讨。首先,在神经元模型的基础上提出了专家预测模型。然后,根据专家预测模型的需要对专家给出的模糊语言判断信息的合成进行了研究,并给出了具体的量化方法。最后,为提高预测方法的精度,提出了一种专家权值的调整方法;算例表明,该方法具有简单实用的特点。

Aiming at the demand of solving complicated prediction problem, this study probes into a method of the experts prediction based on the thought of the artificial neural network. Firstly, a model of the experts prediction is presented on the basis of neural cell model. Then, the integration of the fuzzy judgement information given by the experts is studied for the demand of the prediction model, and a detailed method how to quantify the experts＇ prediction result is presented. Finally, a method of revising the experts＇ weights is put forward for the sake of enhancing the precision of the experts prediction, and an illustrative example is also given to testify the method with simple and applied characteristics.