人工神经网络用于N-甲基氨基甲酸酯类农药急性毒性的预测  

Prediction of toxicity of carbamate pesticide using artificial neural network

作　　者：陈艳[1] 何宇寰 

机构地区：[1]徐州工程学院化学化工学院,江苏徐州221111

出　　处：《分子科学学报》2015年第6期462-467,共6页Journal of Molecular Science

基　　金：国家自然科学基金资助项目(21272095);徐州工程学院培育项目(XKY2011102)

摘　　要：基于拓扑理论计算了N-甲基氨基甲酸酯化合物的电拓扑状态指数(E_n)和分子电性距离矢量(M_k),同时计算了它们理化参数.把2类结构参数用逐步回归分析法进行筛选,确定了4个变量组合E_(31),M_1,M_(60)和SAA.用筛选后的结构参数建立N-甲基氨基甲酸酯农药急性毒性(log1/LD_(50))的神经网络模型,并用模型对log1/LD_(50)值进行了预测,预测结果和实验值比较接近,平均误差为0.071.该方法为开发高效低毒的新型氨基甲酸酯类农药提供理论依据.Based on the topological theory,electrotopological state indexes（En）and electro-negativity distance vector（Mk）of 22 carbamate pesticides were calculated,and meanwhile their physicochemical parameters were calculated.These parameters were selected by stepwise regression method,and four parameters were obtained：E31,M1,M60 and SAA.The neural network model was set up using the preselected parameters and with this model the toxicity（log1/LD50）of carbamate pesticide was predicted.The result showed that the predicted values are close to the experimental values,and the average error is 0.071.This method will provide theoretical support for the design of the novel carbamate pesticides with high efficiency and low toxicity.

关 键 词：拓扑指数 理化参数 神经网络 定量结构-活性相关 

分 类 号：O620[理学—有机化学] TQ450.2[理学—化学]