| Popis: |
Classifcation and regression trees, as well as their variants, are of-the-shelf meth ods 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 optimiza tion algorithms proposed. We illustrate how these powerful formulations enhance the fexibility 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. |
