农药醚菊酯类似物构效关系的人工神经网络方法研究

STRUCTURE-ACTIVITY RELATION OF PYRETHROID COMPOUNDS BY ARTIFICIAL NEURAL NETWORK

本文以醚菊酯类似物作为研究对象，尝试使用神经网络方法进行构效关系分析，并对该种农药活性进行了预测。在所研究的样本集中，由结构预测活性的成功率可达100％，本文的研究表明：神经网络方法以其极强的非线性能力，可望成为农药构效关系研究的一种有效的工具．

Artificial neural network approach was applied to analysis of structure-activity relation of medicine by taking Pyrethroid compounds as an object of study. And the activity of this compound has been predicted. In our research examples, the successful rate reaches 100%. The results indicated that the neural network might be refferred as an effective assistant tchnique for analysis of structure-activity relation of medicine based on ts extremely non-linear ability.