Identifying equations that represent properties in homologous series using structure-structure relations

The Quantitative-Structure-Structure-Property Relationships (QS2PR) technique, which we introduced recently, is adapted to the prediction of properties of pure compounds in homologous series. The QS2PR method involves calculation of the molecular descriptors of a target compound of unknown properties, followed by regression of this vector of molecular descriptors versus a database of compounds with known descriptors and measured properties. The regression model, obtained for the target descriptor in terms of predictive compounds and their coefficients, is then used for predicting properties of the target compound. A structure-structure relationship is derived from the number of carbon atoms and one additional, nonlinearly dependent molecular descriptor of the target compound and three predictive compounds. It is shown that such a relationship can provide predictions of satisfactory precision for many properties of the members of the homologous series. This enables users, without access to libraries of molecular descriptors, to employ the QS2PR technique for property prediction.