摘要
利用分子路径指数矢量表示烷烃分子结构方法 ,结合多元线性回归算法及反传神经网络算法 ,对烷烃摩尔响应值进行处理 ,获得了比文献更佳的预测效果 ,交互校验的相关系数达 0 96
A new method based on a novel molecular topological index vector, called the molecular path vector (MPV), of alkane molecules is proposed and employed for estimation and prediction of the molar response values of various alkanes The novel MPV, p=(P 1,P 2,P 3,P 4,P 5,P 6,P 7,P 8,P 9,P 10 )′, which derived directly from the interaction terms of molecular graph, is used to characterize well molecular structures of all alkanes from one through ten or eleven carbon atoms It showed that there exists very good correlation between the MPV elements and molar response values on both FID and TCD detectors in classical gas chromatography Based on the given calibration set with different sample numbers and by using the practical multiple linear regression, the quantitative structure response relationship (QSRR) equations, for the molar response values ( S M) on both FID and TCD, are respectively given as follows:\; S\- M(FID)= 15 400 488 1 + 17 990 599 5 X 1- 0 165 211 6 X 2- 0 697 410 3 X 3- 0 845 239 0 X 4- 0 267 100 0 X 5- 1 565 727 3 X 6+ 0 094 444 0 X 7, n=50,m=7,r=\{0 997 6\},S T=26 132, S R=1 965 1, E V=99 72%,RMS=1 801, F =1 231 71\; S\- M(TCD)= 11 994 699 6 + 29 149 091 6 X 1- 4 745 166 9 X 2- 3 767 338 5 X 3- 1 494 833 0 X 4- 1 627 883 1 X 5- 0 793 461 1 X 6- 3 056 609 3 X 7, n=32, m=7, r= 0 996 8 , S T=15 72, S R=1 431 0, E V=99 59%,RMS=1 239, F =531 227 where the independent descriptor variables, X 1 X 7 , refer to the elements, P 1,P 2,P 3,P 4,P 5,P 6,P 7 in the molecular path vector for all samples in both FID and TCD training sets; n,r, S T, S R, E V, RMS and F are the sample number, regression coefficient, total standard deviation, standard residual deviation, explained variance, rooted mean squared error and F statistic value, respectively To test both models by using back propagation neural network (BPNN) with the topological structure NN(7 4 2) and the cross validation through leave one out (LOO) procedure, the correlation coefficient of cross validation is over 0 96 Because there exists a quite good linear relationship between the molar responses and molecular path parameters, BPNN ( r =0 989 and 0 968) does not show its nonlinear advantage over multiple linear regression(MLR) ( r = 0 997 6 and 0 996 8 ) in both presently examined cases, FID and TCD in the GC technique, for molecular modelling and quantitative prediction.
出处
《色谱》
CAS
CSCD
北大核心
2000年第6期480-486,共7页
Chinese Journal of Chromatography
基金
国家教委霍英东基金
国家"春晖计划"教育部启动基金
重庆大学研究基金资助项目
关键词
摩尔响应值
分子路径指数矢量
QSRR
烷烃
色谱
molar response values
molecular path vector(MPV)
quantitative structure response relationship(QSRR)
neural networks
molecular modelling
