| Abstract
| - To date, most reported quantitativestructure−property relationship (QSPR) methods to predict vaporpressurerely on, at least, some empirical data, such as boiling points,critical pressures, and critical temperatures.This limits their usefulness to available chemicals and incurs thetime and expense of experimentation. Amodel to predict vapor pressure from only computationally derivedmolecular descriptors, allowing studyof hypothetical structures, is described here. Several multilinearregressions and artificial neural networkanalyses were tested with a range of descriptors (e.g., topological andquantum mechanical) derived solelyfrom computations on molecular structure data. From a set of 479compounds, a linear regression with anr2 of 0.960 was achieved using polarizibilityand polar functional group counts as descriptors. Thisnewcomputationally based model also proves to be more accurate and worksover a wider range of compoundclasses than most previously reported models.
|
