Abstract

The most important domain part of chemometrics is the studies which relate the properties of molecules to their structural characteristics. In this work we use the quantitative structure – property relationship (QSPR) for providing a linear model for predicting the heat capacity of alcohols and aldehydes in liquid phase at 298 K. To perform this research, a set of 32 alcohol and aldehyde as data series was selected then topologic, electronic and geometric descriptors for data series was calculated. Finally, multiple linear regression method was used for selecting more important descriptors and obtaining convenient model. 

Key words: Quantitative structure – property relationship, heat capacity, descriptor, multiple linear regressions, cross validation test.