论文部分内容阅读
运用人工神经网络自组织神经树方法研究了氢醌类化合物的抑菌活性与其结构特征参数间的非线性关系，并对一个“未知样本”的活性类别作了预测，结果与实际一致，表明该方法可作为氢醌类药物构效关系研究的有效辅助手段。
The nonlinear relationship between the antibacterial activities of hydroquinone compounds and their structural parameters was studied by means of artificial neural network and self-organizing neural tree method. The prediction of the activity category of a “unknown sample” was consistent with the actual results. The method can be used as an effective adjunct to study the QSAR of hydroquinone drugs.