人工神经网络及分子拓扑参数在酚类有机物QSBR研究中的应用

An Application of Artificial Neural Networks and MolecularTopological Index for the QSBR of Phenolic Organics*

利用分子拓扑参数作为输入参数，探索了人工神经网络对２７种酚类有机物的定量结构－生物降解性能关系（ＱＳＢＲ）．结果表明，将人工神经网络运用于有机物的生物降解性能建模是可行的．所建模型预测结果和文献数据十分接近，预测能力优于已有文献报道。

A quantitative structurebiodegradability relationships (QSBR)type model using artificial neural networks (ANN)was established for the 27 phenolic compounds, in which molecular topological index are calculated and taken as the input parameters. The results show that the model developed can make a better agreement between predicted and observed values for the biodegradability of the tested compounds than ever before.