基于样条偏最小二乘法的中药量效关系研究  

Research on TCM Dose-Effect Relationship Based on Spline Partial Least Squares Method

作　　者：李欢 万珊 聂斌[1] LI Huan;WAN Shan;NIE Bin(College of Computer Science,Jiangxi University of Chinese Medicine,Nanchang 330004,China)

机构地区：[1]江西中医药大学计算机学院,江西南昌330004

出　　处：《现代信息科技》2023年第7期131-133,137,共4页Modern Information Technology

基　　金：江西省教育厅科技计划研究项目(GJJ211256);江西省大学生创新创业训练计划项目(S202110412057)。

摘　　要：中药量效之间呈现多成分、多靶点的非线性关系,偏最小二乘回归是一种线性回归方法,不能充分表达中药量效之间的非线性关系。基于此,文章建立基于样条偏最小二乘的中药量效关系模型,该模型先对自变量进行三次B样条变化,再进行偏最小二乘回归建模。分别在6组UCI数据集、2组中药数据上展开实验,结果表明,基于样条偏最小二乘回归模型能够很好地拟合非线性数据,并且对中药量效关系研究可行有效。There is a nonlinear relationship between the dosage and effect of Traditional Chinese Medicine(TCM)with multi components and multi targets,and Partial Least Squares Regression(PLSR)is a linear regression that does not adequately express the nonlinear relationship between dosage and effect of TCM.Based on this,this paper establishes a dose-effect relationship model for TCM based on spline partial least squares.The model firstly performs three B-spline changes on the independent variables,and then conducts PLSR modeling.Experiments are carried out on 6 sets of UCI datasets and 2 sets of TCM datasets.The results show that the spline PLSR model can fit the nonlinear data well,and it is feasible and effective for the study of the dose-effect relationship of TCM.

关 键 词：量效关系 非线性 中医药信息学 偏最小二乘 样条函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]