芬太尼衍生物构效关系的自组织人工神经网络研究

Studies on Structure-activity Relationship of Fentanyl Derivatives by T. Kohonen Self-organization Model

本文运用Ｔ．Ｋｏｈｏｎｅｎ自组织人工神经网络建立了芬太尼衍生物的镇痛活性与其结构特征参数间的非线性关系．结果表明，该方法识别成功率较高，可望成为药物构效关系研究的有效辅助手段．

In this papper, the non-linear relationship between the characteristic parameters and analgesic activities of a series of fentanyl derivatives is calculated by T. Kohonen self-organization model. The results show that the performance of the neural network is good and the successful classification rate is high. Therefore, the model might be used as an effective assistant technique for the investigation of structure-activity relationship of drugs.