SELECTION OF DECISION BOUNDARIES FOR LOGISTIC REGRESSION
Access this Article
Search this Article
We propose a method that selects decision boundaries for the logistic regression model by applying sparse regularization. We can investigate which decision boundaries are truly necessary for the multinomial logistic regression model by letting some of the coefficient parameters or the differences between them approach zero. The model is estimated by the maximum penalized likelihood method with a fused lasso-type penalty. We also introduce various model selection criteria for evaluating models estimated by the penalized likelihood method. Simulation studies are conducted in order to evaluate the effectiveness of the proposed method. Real data analysis provides new insights into how each of the predictors contributes to the classification.
- Bulletin of informatics and cybernetics
Bulletin of informatics and cybernetics (47), 83-95, 2015-12