EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3687. On the optimization of risk scores for continuous predictors
Invited abstract in session WD-27: Machine Learning for and with Mathematical Optimization, stream Mathematical Optimization for XAI.
Wednesday, 14:30-16:00Room: 047 (building: 208)
Authors (first author is the speaker)
| 1. | Cristina Molero-Río
|
| École Polytechnique | |
| 2. | Claudia D'Ambrosio
|
| LIX, CNRS - Ecole Polytechnique |
Abstract
In this talk, we propose a novel Mixed-Integer Non-Linear Optimization formulation to construct a risk score. A trade-off between prediction accuracy and sparsity is sought. Previous approaches are typically designed to handle binary datasets, where numerical predictor variables are discretized in a preprocessing step by using arbitrary thresholds, such as quantiles. In contrast, we allow the model decide for each continuous predictor variable the particular threshold that is critical for prediction. The resulting optimization problem is tested in synthetic and real-world datasets.
Keywords
- Machine Learning
- Programming, Mixed-Integer
- Programming, Nonlinear
Status: accepted
Back to the list of papers