• 关键词：人工神经网络  芳基脲  土壤吸附系数  定量结构活性关系(QSAR)

• 基金项目：

• 刘够生
• 华东理工大学资源与环境工程学院, 上海 200237

• 宋兴福
• 华东理工大学资源与环境工程学院, 上海 200237

• 于建国
• 华东理工大学资源与环境工程学院, 上海 200237

• 钱旭红
• 华东理工大学资源与环境工程学院, 上海 200237

• 摘要：对38种芳基脲类化合物的土壤吸附系数进行了定量结构活性相关研究.利用量子化学的AM1方法提取化合物的结构参数, 运用多元回归分析和人工神经网络对该化合物进行比较研究.研究结果表明, 神经网络法所得结果优于多元回归分析结果, 并对分析结果进行了吸附机理讨论.

• Abstract：Soil sorption coefficient predictions of substituted phenylureas were performed by multiple linear regression (MLR) and artificial neural networks (ANN) Molecular structural descriptors were calculated by AM1 quantum chemical computation. Quantitative structure activity relationships (QASR) models were obtained. The results showed that ANN owned a better prediction ability than MLR. The soil sorption mechanism was discussed in detail in this paper.

• Key words：artificial neural networks  substituted phenylureas  soil sorption coefficients  quantitative structure activity relationships (QSAR)

• 
  摘要点击次数： 1933 全文下载次数： 3039