Electronegativity topologic index (ETI) method has been proposed for the numerization of structuxal information. The use of group electronegativity (ETIO) in ETI is more reasonable than that of the intrinsic state term used in the E-state model. The ETI method can be used for direct description of molecular structure information in atom level. Application example showed that the ETI method in combination with artificial neural network and regression to build QSAR model was suitable for some systems.
