QSPR Modeling of Lipophilicity by Means of Correlation Weights of Local Graph Invariants

A quantitative structure–property modeling of the log P (octanol/water partition coefficient) of 76 industrial chemicals is presented. Estimations are performed by means of correlation weighting of local invariants of labeled hydrogen–filled graphs. Results are quite satisfactory, with lower average deviations than other calculations performed with similar theoretical methods. Some possible applications and further extensions of the computation procedure to estimate other physico–chemical or biological properties are mentioned.