Improvement of Artificial Neural Networks and Its Applications to QSARs.

人工神经网络的改进及其在 QSAR 中的应用。

• 批准号: 08672476
• 负责人: TAKAGI Tatsuya
• 金额: $ 1.28万
• 依托单位: Osaka University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 1996
• 资助国家: 日本
• 起止时间: 1996 至 1997
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-08672476/
• 关键词: QSAR; Neural Networks; Robust Estimation; Computer Statistics; Shift Test Method; 医薬品分子設計

In this year, we carried out the quantitative structure activity relationship analyzes of the 2,4-diamino-6,6-dimethyl-5-phynylhydrotriazine derivatives, which show the inhibition activities for dihydrofolatereductase using the previous results about the robust artificial neural network method.Firstly, we adjusted the number of neurons in the hidden layr by extended shift test method shifting the teacher signals.The results show that the best number of the neurons in the hidden layr was 25-30 because the good relations between the parent point and the background points were found by the extended shift test method. Therefore, we used the number of neurons in hidden layrs through out this project. The extended shift test method for the input descriptors (pi_2, pi_3, pi_4, MR_2, MR_3, MR_4, SIGMAsigma_<3,4>) showed that there was no deleted one.We used the two kinds of robust artificial neural network techniques.1) Firstly, the back propagation learning were carried out using the 90% linearity, and then, the 38 data which showed the large errors were deteted from the learning data. Finally, normal artificial neural networks (0% linearity) were adopted.2) The linearity was stepwisely decreased from 80% to 0% and the weight, which showed the large errors, was decreased in each step.In the case of 1), the residual sum of squares resulted in E=23.27. The predictivity of the artificial neural networks was remarkably improved compared with E=30.01 when not using the robust techniques. And in the case of 2), the final root mean residual sum of squares, Ep (=0.49) showed the remarkable improvement compared with Ep=0.69 when not using the robust artificial neural networks.Otherwise, we carried out the classifying of bioactive chemical substances using livingstone-type 5-layred artificial neural networks and the good results were obtained. These results were reported in the Synposium on Chemical Information and Computer Sciences at Kumamoto, 1997.

今年，我们利用鲁棒人工神经网络方法的结果，对2,4-二氨基-6,6-二甲基-5-苯基氢三嗪衍生物进行了定量结构活性关系分析，显示了对二氢叶酸还原酶的抑制活性。 结果表明，由于扩展平移测试方法发现父点和背景点之间存在良好的关系，因此隐含层神经元的最佳数量为25-30个。因此，我们在整个项目中使用了隐藏层中神经元的数量。对输入描述符(pi_2,pi_3,pi_4,MR_2,MR_3,MR_4,SIGMAsigma_<3,4>)的扩展移位测试方法显示没有删除。我们使用了两种鲁棒的人工神经网络技术。1)首先使用90%线性度进行反向传播学习，然后对表现出较大误差的38个数据进行反向传播学习。 从学习数据中检测出来。最后，采用普通的人工神经网络（0％线性）。2）线性从80％逐步降低到0％，并且每一步都减少显示出较大误差的权重。在1）的情况下，残差平方和导致E=23.27。与不使用稳健技术时的 E=30.01 相比，人工神经网络的预测能力显着提高。而在情况2）中，最终的均方根残差平方和Ep（=0.49）与不使用鲁棒人工神经网络时Ep=0.69相比显示出显着的改善。另外，我们使用利文斯通型5层人工神经网络进行生物活性化学物质的分类，并获得了良好的结果。这些结果发表在 1997 年熊本化学信息和计算机科学研讨会上。

项目成果

Takashi MASUDA: "Introduction of Slovent-Accessible Surface Area in the Calculation of the Hydrophobicity Parameter logP from an Atomistic Approach." Journal of Pharmaceutical Sciences. vol.86(印刷中). (1997)

Takashi MASUDA：“通过原子方法计算疏水性参数 logP 的斯洛文可及表面积简介”，《药物科学杂志》第 86 卷（出版中）。

T.Masuda et al.: "Introduction of Solvent-Accessible Surface Area in the Calculation of the Hydrophobicity Parameter logp from an Atomistic Approach" J.Pharm.Sci.86(1). 57-63 (1997)

T.Masuda 等人：“在原子方法计算疏水性参数 logp 中引入溶剂可及表面积”J.Pharm.Sci.86(1)。

高木達也: "計算機統計学の薬学、微生物学への応用" 防菌防黴. 24(1). 39-48 (1996)

Tatsuya Takagi：“计算机统计在药理学和微生物学中的应用”抗菌和抗真菌24（1）（1996）。

Masuda, T. ; Jikihara, T. ; Nakamura, K. ; Kimura, A. ; Takagi, T. ; Fujiwara, H.: "Intruduction of Solvent Accessible Surface Area in the Calculation of the Hydrophobicity Parameter logP from an Atomic Approach" J.Pharm.Sci.vol.86, No.1. 57-63 (1997)

增田，T.；

高木 達也: "計算機統計学の薬学、微生物への応用" 防菌防微. 24(1). 39-48 (1996)

Tatsuya Takagi：“计算机统计在制药科学和微生物中的应用”Antibacterial Antimicrobial 24(1) (1996)。