Studies On The Consensus Strategy Was Used In The Modeling Of 3D-QSAR And HQSAR

In chemometric studies,it is often necessary to analyze substances by establishing relevant mathematical models.Therefore,a variety of modeling methods for analytical chemistry were born.However,a single model is inherently susceptible to the difficulties associated with data quality and sample number.In this dissertation,I take the physical and chemical properties of hydrocarbons as the research object,consensus strategy was used in the modeling of 3D-QSAR and HQSAR,including the following works:The first chapter summarizes the basic principles of 3D-QSAR,HQSAR and integrated modeling,focusing on the basic theory of integrated modeling and modeling methods of consensus modeling.In the second chapter,the quantitative relationship between the refractive index and structure of hydrocarbons was modeled by using the methods based on HQSAR and 3D-QSAR.Leave-one-out cross validation and external validation were carried out to assess the prediction ability of the developed models.All the developed models can be used to predict the refractive index of hydrocarbons.The results show that the Co MFA and HQSAR model can be used to predict the refractive index of hydrocarbons,but the HQSAR model is better than Co MFA model.The Co MSIA model cannot be used to predict.In the chapter three,the quantitative relationship between the refractive index and structure of hydrocarbons was modeled by using the consensus strategy was used in the modeling of 3D-QSAR.Using the bagging method to perturb the samples in the training set,and then 3D-QSAR method is used to establish multiple integrated modeling member models.Screening models with better predictive ability,and finally combining the results of our 3D-QSAR integrated modeling results through consensus principles.Compared with the results of the traditional 3D-QSAR modeling method in chapter 2,the consensus strategy is indeed helpful to improve the predictive ability of 3D-QSAR modeling.In the chapter four,the quantitative relationship between the refractive index and structure of hydrocarbons was modeled by using the consensus strategy was used in the modeling of HQSAR.The integrated modeling steps are roughly the same as in Chapter 3,except that the HQSAR method is used to build the member model.Compared with the results of the traditional HQSAR modeling method in chapter 2,the consensus strategy is indeed helpful to improve the predictive ability of HQSAR modeling.In the chapter five,the quantitative relationship between the boiling point and structure of hydrocarbons was modeled by using the methods based on HQSAR and 3D-QSAR.Leave-one-out cross validation and external validation were carried out to assess the prediction ability of the developed models.All the developed models can be used to predict the boiling point of hydrocarbons.The results show that the Co MFA and HQSAR model can be used to predict the boiling point of hydrocarbons,but the HQSAR model is better than Co MFA model.In the chapter six,the quantitative relationship between the boiling point and structure of hydrocarbons was modeled by using the consensus strategy was used in the modeling of 3D-QSAR.Compared with the results of the traditional 3D-QSAR modeling method in chapter five,the consensus strategy is still helpful to improve the predictive ability of 3D-QSAR modeling.In the chapter seven,the quantitative relationship between the boiling point and structure of hydrocarbons was modeled by using the consensus strategy was used in the modeling of HQSAR.Compared with the results of the traditional HQSAR modeling method in chapter five,the consensus strategy is helpful to improve the predictive ability of HQSAR modeling.But there was no significant improvement in predictive ability.