Mathematical optimization in classification and regression trees

Mathematical optimization in classification and regression trees
复制标题

DOI:
10.1007/s11750-021-00594-1
复制
发表时间:
2021-03-17
期刊:
TOP
影响因子:
1.7
通讯作者:
Romero Morales D
Romero Morales D
中科院分区:
管理学4区
文献类型:
--
作者:
Carrizosa E;Molero-Río C;Romero Morales D

文献摘要

参考文献

被引文献

相似文献

分类和回归树及其变体是机器学习中的现成方法。在本文中,我们回顾了最近的贡献内的连续优化和混合线性优化范式开发新的配方在这一研究领域。我们比较的决策变量的性质和所需的约束条件,以及提出的优化算法。我们说明了这些强大的配方如何增强树模型的灵活性,更适合于将理想的属性,如成本敏感性,可解释性和公平性,并处理复杂的数据,如功能数据。
Classification and regression trees, as well as their variants, are off-the-shelf methods in Machine Learning. In this paper, we review recent contributions within the Continuous Optimization and the Mixed-Integer Linear Optimization paradigms to develop novel formulations in this research area. We compare those in terms of the nature of the decision variables and the constraints required, as well as the optimization algorithms proposed. We illustrate how these powerful formulations enhance the flexibility of tree models, being better suited to incorporate desirable properties such as cost-sensitivity, explainability, and fairness, and to deal with complex data, such as functional data.
DOI: 10.1016/j.ejor.2017.02.046
发表时间: 2017-10-01
影响因子: 6.4
作者:
Athanasopoulos, George;Hyndman, Rob J.;Petropoulos, Fotios
通讯作者: Petropoulos, Fotios
DOI: 10.1007/s10994-017-5633-9
发表时间: 2017-07-01
期刊: MACHINE LEARNING
影响因子: 7.5
作者:
Bertsimas, Dimitris;Dunn, Jack
通讯作者: Dunn, Jack
DOI: 10.1007/s11749-016-0481-7
发表时间: 2016-06-01
期刊: TEST
影响因子: 1.3
作者:
Biau, Gerard;Scornet, Erwan
通讯作者: Scornet, Erwan
DOI: 10.1287/mnsc.49.3.312.12739
发表时间: 2003-03-01
期刊: MANAGEMENT SCIENCE
影响因子: 5.4
作者:
Baesens, B;Setiono, R;Vanthienen, J
通讯作者: Vanthienen, J
DOI: 10.1016/j.omega.2019.05.004
发表时间: 2020-10-01
影响因子: 6.9
作者:
Benitez-Pena, Sandra;Bogetoft, Peter;Morales, Dolores Romero
通讯作者: Morales, Dolores Romero