Layered Rule Induction

This paper proposes a new framework for rule induction methods, called “layered rule induction”, based on rule layers con- strained by inequalities of statistical indices, such as confidence and support. The change of indices with an additional example reflects their sensitivity, and four patterns should be considered if confidence and support are focused on. Then, by using these two pairs of inequalities obtained by analysis, the proposed method classifies a set of formulae into four layers: the rule layer, subrule layer (in and out) and the non-rule layer. Using these layers, updates of probabilistic rules are equivalent to their move- ment between layers. Rules can be extracted from each rule layer. The proposed method was evaluated on datasets regarding headaches and meningitis, and the results show that the proposed method outperforms the conventional methods