Group training methods, described

In the past, a few methods for estimating the liquid viscosity have been proposed, like the group contribution method of Orrick/Erbar [32], Sastri and Rao [69, 70], or van Velzen et al. [71]. They can hardly be recommended, as their application is not easy and their results are often doubtful. For instance, the van Velzen method, often regarded as the most reliable one, produced a relative mean deviation of 92.8% in a benchmark with 670 components [72]. Although data points for the liquid viscosity often widely scatter and deviations can be easier tolerated than for the thermal properties, this value indicates that a lot of components, probably those which were not in the training set, are described poorly. 

Another group has evaluated self-organizing maps [63] and shape/ pharmacophore models [64]. They developed a new method termed SQUIRREL to compare molecules in terms of both shape and pharmacophore points. Thus from a commercial library of 199,272 compounds, 1926 were selected based on self-organizing maps trained on peroxisome proliferator-activated receptor a (PPARa) "activity islands." The compounds were further evaluated with SQUIRREL and 7 out of 21 molecules selected were found to be active in PPARa. Furthermore, a new virtual screening technique (PhAST) was developed based on representation of molecules as text strings that describe their pharmacophores [65]. 

Previous chapters have described simple methods of identification, and generally these are sufficient in order to assign an unknown sample to a certain group of plastics. Of course, with these simple methods, especially in dealing with plastics of complicated composition, one can only obtain some qualitative information. In order to obtain more detailed information it is necessary to use more advanced methods of analysis (Table 8.1), which can only be carried out by specially trained personnel. 

---