Group contribution methods are the methods of molecular
property estimation from the information of groups which construct the molecule.
We have selected the group contribution method as the
input of the neural network. This selection means that users can estimate
molecular properties by using only the information of the type and number
of groups which construct the molecule. In addition, other information obtained
from two dimensional structural information, such as structural isomer information
(cis and trans etc.), is given as necessary. The additional information about
a molecule will increase the accuracy of property estimation.
HOME
Previous Page
TOP