Generalized Linear Rule Models
[edit]
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:66876696, 2019.
Abstract
This paper considers generalized linear models using rulebased features, also referred to as rule ensembles, for regression and probabilistic classification. Rules facilitate model interpretation while also capturing nonlinear dependences and interactions. Our problem formulation accordingly trades off rule set complexity and prediction accuracy. Column generation is used to optimize over an exponentially large space of rules without pregenerating a large subset of candidates or greedily boosting rules one by one. The column generation subproblem is solved using either integer programming or a heuristic optimizing the same objective. In experiments involving logistic and linear regression, the proposed methods obtain better accuracycomplexity tradeoffs than existing rule ensemble algorithms. At one end of the tradeoff, the methods are competitive with less interpretable benchmark models.
Related Material


