基于特征提取的中药水提液膜分离预测系统

英文篇名：Forecasting system of Chinese medicine’s water extract in membrane separation process based on feature selection
中文刊名：计算机工程与设计
英文刊名：Computer Engineering and Design
作者：李玲娟 ; 郭立玮
英文作者：LI Ling-juan1 ; GUO Li-wei2+ (1. College of Information and Technology ; Nanjing University of Chinese Medicine ; Nanjing 210046 ; China ; 2. Research Institute of Chinese Medicine ; Nanjing University of Chinese Medicine ; Nanjing 210046 ; China)
中文关键词：特征提取 ; 遗传算法 ; 神经网络 ; 支持向量机 ; 模型预测系统
英文关键词：feature selection ; genetic algorithm ; artificial neutral network ; support vector machine ; mod el forecasting system
出版日期：2010-05-16
机构：南京中医药大学信息技术学院;南京中医药大学中医药研究院;
年：2010
期：09
出版单位：计算机工程与设计

摘要

为了找出中药水提液膜过程中影响膜污染的主要因素和预测膜污染的程度以防止膜污染,提出了应用遗传神经网络提取影响中药水提液膜分离过程的特征因素的方法,并以特征因素为输入向量,使用神经网络、支持向量机等建立预测模型,开发并实现了集成化的综合分析和预测系统。介绍了中药水提液膜分离预测系统的体系结构、主要功能、运行情况及开发的关键技术。实验结果表明,该集成化的综合分析系统较单一分析建模预测精度更高。
With the purpose of finding the critical factors which exert great influences on membrane pollution in the filtration of Chinese medicine’s water extract and preventing membrane pollution by forecasting the pollution degree, an account of how to select features in membrane separation process by genetic algorithm and artificial neutral network, and how to build up a forecasting model by artificial neutral network and support vector machine, taking selected features as input vector. With characterization data of membrane separation of Chinese medicine’s water extract, by means of module development and system integration. Aimed at developing a forecasting system whose systematic structure, main function, operation situation and key developing technology of are also discussed, which can predict and analyze membrane fouling in membrane separation process of Chinese medicine’s water extract. Besides, the finding indicated that this integrated analysis system is superior to the mono-analysis in forecasting accuracy of modeling building.

引文

[1]YongSeog Kim,Nick Street W,Filippo Menczer.Feature selec-tion in data mining[D].University of Iowa,2008:1-22.
    [2]Isabelle Guyon.An introduction to variable and feature selection[J].Journal of machine research,2005(3):1157-1182.
    [3]王世香.精通Matlab接口与编程[M].北京:电子工业出版社,2007:279-284.
    [4]葛哲学,孙志强.神经网络理论与MATLAB R2007实现[M].北京:电子工业出版社,2007:25-87.
    [5]雷英杰.MATLAB遗传工具箱及应用[M].西安:西安电子科技大学出版社,2005:11-31.
    [6]齐珺.基于遗传-支持向量机和遗传-径向基网络的有机物正辛醇-水分配系数QSPR研究[J].环境科学,2008,29(1):212-218.
    [7]乔维德.遗传算法和神经网络在交通事故预测中的应用[J].电气传动自动化,2008,30(1):41-44.
    [8]齐珺.基于遗传-支持向量机和遗传-径向基网络的有机物正辛醇-水分配系数QSPR研究[J].环境科学,2008,29(1):212-218.
    [9]陈一超.基于遗传神经网络的地震预测研究[J].计算机应用与软件,2008(4):135-137.
    [10]邓乃杨,田硬杰.数据挖掘中的新方法——支持向量机[M].北京:科学出版社,2005:1-342.
    [11]李盼池,许少华.支持向量机在模式识别中的核函数特性分析[J].计算机工程与设计,2005,26(2):302-304.
    [12]范英飚,潘妍.基于支持向量机自适应核的改进算法[J].计算机工程与设计,2008,29(12):6073-6111.
    [13]朱树先,张仁杰.支持向量核函数选择的研究[J].科学技术与工程,2008,8(16):4513-4517.