Data mining for seeking accurate quantitative relationship between molecular structure and GC retention indices of alkanes by projection pursuit

Primary data mining on alkanes for seeking accurate quantitative relationship between molecular structure and retention indices of gas chromatography is developed in this paper. Based on the results obtained from projection pursuit (PP), a new variable named class distance variable, which essentially describes the branching structure of the alkanes, is proposed. With the help of the new variable, both fitting and prediction accuracy of the regression model can be dramatically improved. The results obtained in this work show that the technique of PP developed in statistics is a quite promising tool for seeking accurate quantitative structure-activity relationship (QSAR) and/or quantitative structure-property relationship (QSPR) researches.     

Topological indices and QSPR analysis of some chemical structures applied for the treatment of heart patients

In this study, various beta-blocker drugs used for heart disease were analyzed, and their degree-based topological indices derived from the M-polynomial were calculated. Linear and quadratic regression analysis was used to obtain quantitative structure-property relationship models between the topological indices and eight the physicochemical properties of the drugs to determine their effectiveness. The results show that the harmonic index was the best predictor for boiling point, flashpoint, and enthalpy of vaporization, while the redefined third Zagreb index was effective for polarizability, molar refractivity, and molar volume. The inverse sum indeg index was found to be effective for molar refractivity, and the second modified Zagreb index was surface tension in linear regression models. In addition, the redefined third Zagreb index was the best predictor for polarizability and molar refractivity, while the second modified Zagreb index was effective for molar volume. The SDD index was found to be effective for surface tension in quadratic regression models.