Numerical recipes: the art of scientific computing
Numerical recipes: the art of scientific computing
On the Optimality of the Simple Bayesian Classifier under Zero-One Loss
Machine Learning - Special issue on learning with probabilistic representations
Machine Learning
The Journal of Machine Learning Research
Not So Naive Bayes: Aggregating One-Dependence Estimators
Machine Learning
Sparse Multinomial Logistic Regression: Fast Algorithms and Generalization Bounds
IEEE Transactions on Pattern Analysis and Machine Intelligence
Compact approximations to Bayesian predictive distributions
ICML '05 Proceedings of the 22nd international conference on Machine learning
| Hi-index | 0.01 |
We present a learning algorithm for nominal vector data. It builds a complex classifier by adding iteratively a simple function that modifies the current classifier. In order to limit overtraining problem we focus on a class of such functions for which optimal Bayesian learning is tractable. We investigate a few classes of functions that yield to models that are similar to Nai@?ve Bayes and logistic classification. We report experimental results for a collection of standard data sets that show that our learning algorithm outperforms standard learning of such these standard models.