| 改进型BP神经网络用于有机物毒性的QSAR研究 |
| |
| 引用本文： | 束庆海,何池洋. 改进型BP神经网络用于有机物毒性的QSAR研究[J]. 安庆师范学院学报(自然科学版), 2002, 8(3): 32-34. DOI: 10.3969/j.issn.1007-4260.2002.03.012 |
| |
| 作者姓名： | 束庆海 何池洋 |
| |
| 作者单位： | 安庆师范学院,化学系,安徽,安庆,246011 |
| |
| 基金项目： | 安庆师范学院校科研和教改项目;; |
| |
| 摘 要： | 引进改进型自相关拓扑指数,以基于动量-自适应学习率调整算法的BP神经网络原理研究了36种有机污染物对发光菌的半数发光抑制率EC50的定量结构-活性相关关系,建立了相应的网络模型,用于未知有机物毒性的预测,其准确性明显优于传统的多元线性回归法,取得了令人满意的结果.
|
| 关 键 词： | 人工神经网络 有机物 毒性 QSAR |
| 文章编号： | 1007-4260(2002)03-0032-03 |
Quantitative Structure-activity Relationship Studies on the Toxicities of Organic Chemicals Using Modified Back-propagation Neural Network |
| |
| SHU Qing hai,HE Chi yang. Quantitative Structure-activity Relationship Studies on the Toxicities of Organic Chemicals Using Modified Back-propagation Neural Network[J]. Journal of Anqing Teachers College(Natural Science Edition), 2002, 8(3): 32-34. DOI: 10.3969/j.issn.1007-4260.2002.03.012 |
| |
| Authors: | SHU Qing hai HE Chi yang |
| |
| Abstract: | The quantitative structure activity relationship of the toxicities (EC 50 of photobacterium phosphoreum) of 36 kinds of organic pollutions is studied,based on modified self correlated topologic index,and back propagation neural network is used based on momentum and self adjusted learning rate algorithm.A network model is produced with a better accuracy in the prediction to the toxicities of unknown organic chemicals than that of typical multiple linear regression. |
| |
| Keywords: | artificial neural network toxicities quantitative structure activity relationship QSAR |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
