改进共识QSAR建模特征选择的双聚类策略

Q2 Mathematics
María Jimena Martínez , Julieta Sol Dussaut , Ignacio Ponzoni
{"title":"改进共识QSAR建模特征选择的双聚类策略","authors":"María Jimena Martínez ,&nbsp;Julieta Sol Dussaut ,&nbsp;Ignacio Ponzoni","doi":"10.1016/j.endm.2018.07.016","DOIUrl":null,"url":null,"abstract":"<div><p>Feature selection applied to QSAR (Quantitative Structure-Activity Relationship) modeling is a challenging combinatorial optimization problem due to the high dimensionality of the chemical space associated with molecules and the complexity of the physicochemical properties usually studied in Cheminformatics. This derives commonly in classification models with a large number of variables, decreasing the generalization and interpretability of these classifiers. In this paper, a novel strategy based on biclustering analysis is proposed for addressing this problem. The new method is applied as a post-processing step for feature selection outputs generated by consensus feature selection methods. The approach was evaluated using datasets oriented to <em>ready biodegradation</em> prediction of chemical compounds. These preliminary results show that biclustering can help to identify features with low class-discrimination power, which it is useful for reducing the complexity of QSAR models without losing prediction accuracy.</p></div>","PeriodicalId":35408,"journal":{"name":"Electronic Notes in Discrete Mathematics","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2018-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1016/j.endm.2018.07.016","citationCount":"15","resultStr":"{\"title\":\"Biclustering as Strategy for Improving Feature Selection in Consensus QSAR Modeling\",\"authors\":\"María Jimena Martínez ,&nbsp;Julieta Sol Dussaut ,&nbsp;Ignacio Ponzoni\",\"doi\":\"10.1016/j.endm.2018.07.016\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Feature selection applied to QSAR (Quantitative Structure-Activity Relationship) modeling is a challenging combinatorial optimization problem due to the high dimensionality of the chemical space associated with molecules and the complexity of the physicochemical properties usually studied in Cheminformatics. This derives commonly in classification models with a large number of variables, decreasing the generalization and interpretability of these classifiers. In this paper, a novel strategy based on biclustering analysis is proposed for addressing this problem. The new method is applied as a post-processing step for feature selection outputs generated by consensus feature selection methods. The approach was evaluated using datasets oriented to <em>ready biodegradation</em> prediction of chemical compounds. These preliminary results show that biclustering can help to identify features with low class-discrimination power, which it is useful for reducing the complexity of QSAR models without losing prediction accuracy.</p></div>\",\"PeriodicalId\":35408,\"journal\":{\"name\":\"Electronic Notes in Discrete Mathematics\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2018-08-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://sci-hub-pdf.com/10.1016/j.endm.2018.07.016\",\"citationCount\":\"15\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Electronic Notes in Discrete Mathematics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1571065318301604\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"Mathematics\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Electronic Notes in Discrete Mathematics","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1571065318301604","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"Mathematics","Score":null,"Total":0}
引用次数: 15

摘要

由于与分子相关的化学空间的高维性和化学信息学中通常研究的理化性质的复杂性,特征选择应用于QSAR(定量结构-活性关系)建模是一个具有挑战性的组合优化问题。这通常源于具有大量变量的分类模型,降低了这些分类器的泛化和可解释性。本文提出了一种基于双聚类分析的新策略来解决这一问题。将该方法作为共识特征选择方法产生的特征选择输出的后处理步骤。使用面向化合物的现成生物降解预测的数据集对该方法进行了评估。这些初步结果表明,双聚类可以在不影响预测精度的前提下,有效地识别低分类判别能力的特征,从而降低QSAR模型的复杂度。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Biclustering as Strategy for Improving Feature Selection in Consensus QSAR Modeling

Feature selection applied to QSAR (Quantitative Structure-Activity Relationship) modeling is a challenging combinatorial optimization problem due to the high dimensionality of the chemical space associated with molecules and the complexity of the physicochemical properties usually studied in Cheminformatics. This derives commonly in classification models with a large number of variables, decreasing the generalization and interpretability of these classifiers. In this paper, a novel strategy based on biclustering analysis is proposed for addressing this problem. The new method is applied as a post-processing step for feature selection outputs generated by consensus feature selection methods. The approach was evaluated using datasets oriented to ready biodegradation prediction of chemical compounds. These preliminary results show that biclustering can help to identify features with low class-discrimination power, which it is useful for reducing the complexity of QSAR models without losing prediction accuracy.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Electronic Notes in Discrete Mathematics
Electronic Notes in Discrete Mathematics Mathematics-Discrete Mathematics and Combinatorics
CiteScore
1.30
自引率
0.00%
发文量
0
期刊介绍: Electronic Notes in Discrete Mathematics is a venue for the rapid electronic publication of the proceedings of conferences, of lecture notes, monographs and other similar material for which quick publication is appropriate. Organizers of conferences whose proceedings appear in Electronic Notes in Discrete Mathematics, and authors of other material appearing as a volume in the series are allowed to make hard copies of the relevant volume for limited distribution. For example, conference proceedings may be distributed to participants at the meeting, and lecture notes can be distributed to those taking a course based on the material in the volume.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信