摘要
探讨了4个不同温度下有机物的结构与无限稀释活度系数(γ∞)之间的定量关系.以量子化学参数定量刻画有机分子结构,分别采用遗传程序设计(GP)和理论线性溶解能关系(TLSER)构建关联模型,并对两种方法进行了比较.结果表明:在4个不同温度下,无论从标准偏差还是从平方相关系数看,GP所得结果均优于TLSER,并且GP所得回归方程不仅具有形式上的灵活性,而且方程所含变量较少.
The relationships between structure and infinite dilution activity coefficients of organic molecules were studied in four different temperatures, quantum chemical parameters were used to describe the organic molecular structure, correlation equations were established and compared by using genetic programming (GP) and theoretical linear solvation energy relationships (TLSER) respectively. The results showed that GP was superior to TLSER both on square correlation coefficients and standard deviation of error in four different temperatures. In addition, GP was not only versatile in modeling but also employed fewer variants.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2003年第4期444-448,共5页
Computers and Applied Chemistry
基金
河南省自然科学基金(004031200)
河南省教委自然科学基金(98150027)资助项目
