A Study of Probability Estimation Techniques for Rule Learning

摘要: Rule learning is known for its descriptive and therefore com- prehensible classification models which also yield good class predictions. However, in some application areas, we need probability estimates. For dierent models, such as decision trees, a variety of techniques obtaining estimates have been proposed evaluated. so far, there has no systematic empirical study how these can be adapted to probabilistic rules methods aect the probability-based rankings. In this paper apply several basic estimation membership probabilities rules. We eect shrinkage technique merging with those their generalizations.