摘要
为了定量结构 /毒性相关性研究提取了拓扑指数Am ,分子连接性指数mχt量子化学参数及分子体积等 ,同时运用最佳变量子集算法 (LeapsandBounds)进行了变量的压缩和选择 .进而实施了多元间归分析 ,并由所得结果进行了硝基苯类化合物结构与毒性关系的理论解释 .同时还进行了人工神经网络法对于该类化合物毒性的预测 ,其结果明显好于多元回归法 .
In this study,the topological indices Am(Aml,Am2,Am3)and molecular connectivity indices mχ t ,quantum\|chemical descriptors and the volumes of the 35 nitrobenzene compounds were calculated.The algorithm “Leaps and Bounds” was performed for selection of the variables and the multiregression was carried out with satisfactory results.The toxicity mechanisms were discussed.Furthermore,the quasi\|Newton neural networks were emplogyed;the results obtained were much better than that obtained by multiregression.
出处
《环境科学学报》
CAS
CSCD
北大核心
2000年第4期456-460,共5页
Acta Scientiae Circumstantiae
基金
国家自然科学基金! (2 96 770 16 )
