Support vector regression that takes into consideration the importance of explanatory variables

Support vector regression that takes into consideration the importance of explanatory variables
复制标题

考虑解释变量重要性的支持向量回归

DOI:
10.1002/cem.3327
复制
发表时间:
2020
影响因子:
2.4
通讯作者:
Kaneko Hiromasa
Kaneko Hiromasa
中科院分区:
化学3区
文献类型:
--
作者:
Liu Peizhao;Chen Wei;Okazaki Yutaka;Battie Yann;Brocard Lysiane;Decossas Marion;Pouget Emilie;M?ller-Buschbaum Peter;Kauffmann Brice;Pathan Shaheen;Sagawa Takashi;Oda Reiko;Kaneko Hiromasa

文献摘要

参考文献

相似文献

支持向量回归(SVR)能够考虑解释变量X与目标变量X之间的非线性关系,建立具有较高预测精度的回归模型。此外,使用SVR模型为新样本预测的y值可能会超过训练数据中的实际值。然而,由于支持向量回归机中常用的核函数高斯核是基于样本间的欧氏距离,在建立回归模型时无法考虑X的重要性。因此,本研究的重点是X的重要性可以通过随机森林(RF)来计算,并基于此重要性提出了一种新的支持向量回归方法,称为考虑变量重要性的支持向量回归(VI‐SVR)。因为X是根据重要性加权的,所以X的重要性越大,它对预测值的贡献就越大。使用光谱,定量结构-性质关系(QSPR)和定量结构-活性关系(QSAR)数据集的分析证实,VI-SVR的预测准确性优于SVR。VI‐SVR Python代码可在https://github.com/hkaneko1985/dcekit上获得。
Support vector regression (SVR) is able to consider the nonlinear relationship between explanatory variablesXand a target variableyto build a regression model with high predictive accuracy. Additionally,yvalues predicted with SVR models for new samples can exceed the actualyvalues in training data. However, because the Gaussian kernel, which is a kernel function generally used in SVR, is based on the Euclidean distance between samples, it is unable to consider the importance ofXwhen building the regression model. Therefore, in this study, the focus was on the importance ofXthat can be calculated by random forests (RF), and a novel SVR method, called variable importance‐considering support vector regression (VI‐SVR), was proposed based on this importance. BecauseXis weighted based on importance, the greater the importance ofX, the greater its contribution to the predicted value. Analysis using the spectral, quantitative structure–property relationship (QSPR), and quantitative structure–activity relationship (QSAR) datasets confirmed that the predictive accuracy of VI‐SVR was better than that of SVR. VI‐SVR Python code is available at https://github.com/hkaneko1985/dcekit.
用于构建回归模型的宽谱特征选择 (WiSe)
DOI: 10.1016/j.compchemeng.2018.10.005
发表时间: 2019
期刊: Comput. Chem. Eng.
影响因子: --
作者:
Ricardo R. Rendall;Ivan Castillo;Alix Schmidt;S. Chin;Leo H. Chiang;M. Reis
通讯作者: M. Reis
DOI: 10.1016/j.jmrt.2023.02.141
发表时间: 2023-03-04
影响因子: 6.4
作者:
Wang, Qingjuan;He, Zeen;Yang, Congcong
通讯作者: Yang, Congcong
DOI: 10.1002/ep.11747
发表时间: 2014-04-01
影响因子: 2.8
作者:
Abbasi, M.;Abduli, M. A.;Baghvand, A.
通讯作者: Baghvand, A.
从头设计:从模型到分子
DOI: --
发表时间: 2013
期刊:
影响因子: --
作者:
G. Schneider;K. Baringhaus
通讯作者: K. Baringhaus