Abstract
Rule ensembles are designed to provide a useful trade-off between predictive accuracy and model interpretability. However, the myopic and random search components of current rule ensemble methods can compromise this goal: they often need more rules than necessary to reach a certain accuracy level or can even outright fail to accurately model a distribution that can actually be described well with a few rules. Here, we present a novel approach aiming to fit rule ensembles of maximal predictive power for a given ensemble size (and thus model comprehensibility). In particular, we present an efficient branch-and-bound algorithm that optimally solves the per-rule objective function of the popular second-order gradient boosting framework. Our main insight is that the boosting objective can be tightly bounded in linear time of the number of covered data points. Along with an additional novel pruning technique related to rule redundancy, this leads to a computationally feasible approach for boosting optimal rules that, as we demonstrate on a wide range of common benchmark problems, consistently outperforms the predictive performance of boosting greedy rules.
Original language | English |
---|---|
Title of host publication | Proceedings of the 2021 SIAM International Conference on Data Mining (SDM) |
Editors | Carlotta Demeniconi, Ian Davidson |
Place of Publication | Philadephia PA USA |
Publisher | Society for Industrial & Applied Mathematics (SIAM) |
Pages | 351-359 |
Number of pages | 9 |
ISBN (Electronic) | 9781611976700 |
DOIs | |
Publication status | Published - 2021 |
Event | SIAM International Conference on Data Mining 2021 - Online, United States of America Duration: 29 Apr 2021 → 1 May 2021 https://epubs.siam.org/doi/book/10.1137/1.9781611976700 (Proceedings) https://www.siam.org/conferences/cm/conference/sdm21 (Website) |
Conference
Conference | SIAM International Conference on Data Mining 2021 |
---|---|
Abbreviated title | SDM21 |
Country/Territory | United States of America |
Period | 29/04/21 → 1/05/21 |
Internet address |
Keywords
- rule learning
- boosting
- branch-and-bound
- ensemble methods