Example methyl n-pentanoate

We use the MS classifiers from MSclass [324] to extract structural properties. These classifiers were obtained via classification by regression (Subsection 6.1.1). In training the classifiers for the values of the discriminant function, percent quantiles were calculated. As a result, it is possible to give an estimated precision for each prediction made such that classifiers of low precision can be excluded from further consideration. Unfortunately, by doing so the number of structural properties for generation of molecular and structural formulas will be reduced. 

Altogether there are 160 classifiers available for 85 structural properties (Appendix B). For some properties there are up to four different classifiers differing in the descriptors included and/or in the classification methods used. 

Application of the classifiers to spectrum I of Example 8.2 results in 68 classifier answers of at least 95% precision, with five positive and 63 negative classifiers. The MSclass output for the positive results is as follows  

each line corresponds to a classifier result. The first column is the classifier s name, composed of an abbreviation for the structural property and a classifier ID number. Three of the positive answers are for the property methyl ester . In addition, the precision of the prediction is given in column prec, and the last column shows the classification method used. MSclass uses the methods EDA and ANN with radial base functions (RBF). 

The most important result here is that the functional group methyl ester is predicted as present by three out of four classifiers for this substructure. The other positive classifier, non-aromaticity proves rather unproductive in the following. 

---