摘要
基于遗传神经网络的环境质量评价方法是用遗传算法和BP算法相结合的混合算法来训练环境质量评价神经网络预测模型的权值,即先通过遗传学习算法进行全局训练,再用权重调整BP算法进行精确训练,这一算法克服了BP算法收敛速度慢、易陷入局部极小等缺陷,环境质量评价实例证明提高了预测精度。又将三种评价方法的评价结果进行对比,得到该方法的评价结果比传统的专家评价法的评价结果更加准确。因此该方法也为环境质量评价提供了一种新的研究思路和分析方法。
A new method for training Artificial Neural Network（ANN）based environmental quality assessment prediction model is presented. In this method,the genetic algorithm （GA）,a general purple global search algorithm is used to train the neural network prediction model by updating the weights to minimize the error between the network output and the desired output. It overcomes the limitations of the back propagation algorithm in slow convergent rate and getting into local optima. The environmental quality assessment demonstrates that this method improves the prediction precision. Also contrasting the assessment results from three kind of assessment methods, obtaining this method the assessment result is more accurate than the traditional expert assessment method appraisal result. So this method has offered a kind of new thinking of research and analytic method for environmental quality assessment.
出处
《华北科技学院学报》
2006年第1期62-64,共3页
Journal of North China Institute of Science and Technology
关键词
环境质量评价
遗传算法
BP算法
预测精度
遗传神经网络
environmental quality asseasment
genetic algorithm
BP algorithm
prediction precision
genetic neural network
