Study on a QSAR model of the substituted benzene compounds based on artificial neural network
Dawen Gao, Peng Wang · Ha'erbin gongye daxue xuebao · 2005
In building a Quantitative Structure-Activity Relationship(QSAR) model of substituted benzene compounds,a Vertex degree autocorrelation topological index as a structure parameter was screened by an artificial neural network.The prediction ability and quality assessment of various ANN-QSAR models in the process of network screening were also compared.Results showed that it is feasible to screen structure parameters using an artificial neural network.The method reduced the original 24 structure parameters to 5,and the prediction ability and quality of the model were not influenced.So the study not only expedites the operation speed of the network model,but also establishes the foundation for further study of the bio-toxicity mechanism of substituted benzene compounds