量子进化算法在苹果漫透射近红外光谱分析中的应用

为了研究量子进化算法（quantum-inspired evolutionary algorithm,QEA）在苹果漫透射近红外光谱分析中的应用,先用反向间隔偏最小二乘法（backward interval partial least squares,BiPLS）对光谱信息区间初步定位,再采用QEA算法选择波长点,建立糖度预测模型;同时采用遗传算法（geneticalgorithm,GA）选取波长点建立预测模型,并对2种算法的结果进行比较.结果表明：运行GA算法建立的GA-PLS模型变量数为110,校正均方根误差（root mean standard error of calibration,...

Full description

Saved in:
Bibliographic Details
Published in浙江大学学报(农业与生命科学版) Vol. 37; no. 4; pp. 453 - 459
Main Author 李军良 王从庆
Format Journal Article
LanguageChinese
Published 南京航空航天大学自动化学院,江苏南京,210016 2011
Subjects
变量筛选
漫透射
近红外光谱
量子进化算法
量子进化算法
变量筛选
漫透射
近红外光谱
Online AccessGet full text
ISSN1008-9209
DOI10.3785/j.issn.1008-9209.2011.04.015

Cover

More Information
Summary:为了研究量子进化算法（quantum-inspired evolutionary algorithm,QEA）在苹果漫透射近红外光谱分析中的应用,先用反向间隔偏最小二乘法（backward interval partial least squares,BiPLS）对光谱信息区间初步定位,再采用QEA算法选择波长点,建立糖度预测模型;同时采用遗传算法（geneticalgorithm,GA）选取波长点建立预测模型,并对2种算法的结果进行比较.结果表明：运行GA算法建立的GA-PLS模型变量数为110,校正均方根误差（root mean standard error of calibration,RMSEC）为0.582 0,预测均方根误差（root mean standard error of prediction,RMSEP）为0.612 3;运行QEA算法建立的QEA-PLS模型变量数为194,RMSEC为0.492 7,RMSEP为0.526 0.说明量子进化算法用于苹果漫透射近红外光谱分析可有效提高模型预测精度,相比遗传算法表现出更好的寻优能力.
Bibliography:33-1247/S
quantum-inspired evolutionary algorithms; diffuse transmission; near infrared spectroscopy; variable selection
To study the application of quantum-inspired evolutionary algorithm（QEA） in the analysis of near infrared（NIR） diffuse transmission spectroscopy of apples,first,spectroscopy regions were preselected by using the backward interval partial least squares（BiPLS）.Second,the variables were selected with QEA,and QEA-PLS model was built.Meanwhile,genetic algorithm（GA）-PLS model was developed to contrast with QEA-PLS model.After running GA and QEA 10 times separately,the two best models were chosen from the 10 GA-PLS models and the 10 QEA-PLS models.The results showed that the GA-PLS model had 110 variables,with RMSEC（root mean standard error of calibration） of 0.5820,RMSEP（root mean standard error of prediction） of 0.6123,but the QEA-PLS model had 194 variables,with RMSEC of 0.4927,RMSEP of 0.5260.It is concluded that QEA can be used in the analysis of NIR diffuse transmission spectroscopy of apples
ISSN:1008-9209
DOI:10.3785/j.issn.1008-9209.2011.04.015