Boolean Decision Rules via Column Generation
Authors: Sanjeeb Dash, Oktay Gunluk, Dennis Wei
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | A numerical evaluation is presented using 16 datasets, including one from the ongoing FICO Explainable Machine Learning Challenge [1]. |
| Researcher Affiliation | Industry | Sanjeeb Dash, Oktay Günlük, Dennis Wei IBM Research Yorktown Heights, NY 10598, USA {sanjeebd,gunluk,dwei}@us.ibm.com |
| Pseudocode | No | The paper describes the algorithms and formulations in prose and mathematical equations but does not provide a pseudocode block or a clearly labeled algorithm. |
| Open Source Code | No | The paper does not include any explicit statement or link indicating that the authors' source code for the methodology is publicly available. |
| Open Datasets | Yes | Evaluations were conducted on 15 classification datasets from the UCI repository [23] that have been used in recent works on rule set/Boolean classifiers [39, 19, 46, 52]. In addition, we used recently released data from the FICO Explainable Machine Learning Challenge [1]. |
| Dataset Splits | Yes | Test performance on all datasets is estimated using 10-fold stratified cross-validation (CV). For CG, we used an overall time limit of 300 seconds for training and a time limit of 45 seconds for solving the Pricing Problem in each iteration. nested CV was used to select values of C for CG and for AM, BCD to maximize accuracy on each training set. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used for running the experiments. |
| Software Dependencies | Yes | We use a standard mixed-integer programming solver (CPLEX 12.7.1) to which a time limit can be provided. |
| Experiment Setup | Yes | For CG, we used an overall time limit of 300 seconds for training and a time limit of 45 seconds for solving the Pricing Problem in each iteration. In these experiments, CG was given an overall time limit of 120 seconds for each candidate value of C and the time limit for the Pricing Problem was set to 30 seconds. |