基于改进BP网络的染色合格率预测  

Prediction of Eligibility Rate of Dye Based on Improved BP Neural Network

作　　者：李清华[1] 张美凤[1] 

机构地区：[1]上海大学机电工程与自动化学院,200072

出　　处：《微计算机信息》2006年第04X期93-95,共3页Control & Automation

摘　　要：由于多种因素对染色质量的影响是非线性的,本文在分析影响染色质量因素的基础上,提出了基于BP神经网络的染色合格率预测模型。针对传统BP算法的缺陷,本文采用L-M算法改进BP网络。仿真结果表明,利用该模型预测染色合格率是比较准确而且有效的。Because many factors influence the quality of day nonlinearly, this paper presents a BP neural network model to predict the eligibility rate of dye. It analyses the factors of influencing the quality of dye, and gives the predictition model. Aiming at the shortage of the conventional BP algorithm,the BP neural network improved by L-M algorithm is put forward.The simulation result indicates that the model can predict the eligibility rate of dye exactly and effectively.

关 键 词：染色合格率 L-M算法 BP网络 预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]