The model consists of 147 parameters. The 34 fragments reflect known acceptor bond donor groups. 113 correction factors are applied to reflect the effects of aromatic, olefinic, and hetero atom attachments, sec/tert-positions of certain groups, multiple halogens, intra-molecular hydrogen bonds, effects of ring atoms or ring substituents, interaction of groups close to each other, and certain complex structures. The data set for the acceptor capacity B consists of 3113 chemicals with B less then 4.52. There is a rather large number of hetero-atoms and complex compounds, necessary to overcome the shortcomings of the more simple training sets of the existing model.
The tool is ready for use. It is implemented in the software system ChemProp together with an automated check for the applicability domain in terms of the chemical space. It will be published soon.