基因表达式编程在含氧有机化合物定量结构-保留关系中的应用  

Study on QSRR of Oxygen-containing Organic Compounds Based on Gene Expression Programming

作　　者：丁玲[1] 

机构地区：[1]辽宁现代服务职业技术学院,辽宁沈阳110164

出　　处：《辽宁化工》2016年第4期402-405,共4页Liaoning Chemical Industry

摘　　要：基因表达式编程(GEP)是一种新颖的遗传算法,是一种高度有效、稳定的随机搜索方法。采用GEP对一系列含氧有机化合物的气相色谱保留指数建立定量结构-保留关系(QSRR)的模型,并与人工神经网络(ANN)预测结果进行比较。GEP和ANN在OV-1固定相上,相关系数R分别为0.9908,0.9892;在SE-54上,相关系数R分别为0.9956,0.9891。结果表明:GEP建立的模型优于ANN,具有良好的拟合度和预测精度。Gene expression programming（GEP） is a novel genetic algorithm, a highly effective and stable random searching method. GEP was used to establish models of quantitative structure-retention relationship（QSRR） for a series of oxygen-containing organic compounds GC retention index, and its predictive results were compared with that of artificial neural network（ANN）. The correlation coefficient on OV-1 was 0.9919 and 0.9911 for GEP and ANN respectively; that on SE-54 was 0.9955 and 0.9892.The results show that the GEP predicted results are better than those of ANN.

关 键 词：基因表达式编程 含氧有机化合物 人工神经网络 定量结构-保留关系 

分 类 号：TQ028[化学工程]