Issue No. 02 - March/April (2007 vol. 22)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MIS.2007.19
Jos? Ranilla , University of Oviedo
Luis J. Rodr?guez-Mu? , University of Oviedo
Decision-tree induction and rule-learning methods have proved efficient for concept-learning and data-mining tasks. Modifications to successful algorithms learned from crisp data enable them to deal with cognitive uncertainties that, in general, use entropy as the measurement to select relevant characteristics for the learning task. A proposed heuristic approach induces rules from fuzzy databases, supported by an extension to the fuzzy case of a classical case metric, called the impurity level. An illustrative example tests the algorithm with some data sets and compares it with similar systems.
knowledge acquisition, machine learning, rule-based processing, uncertainty, fuzzy sets, and probabilistic reasoning
L. J. Rodr?guez-Mu? and J. Ranilla, "A Heuristic Approach to Learning Rules from Fuzzy Databases," in IEEE Intelligent Systems, vol. 22, no. , pp. 62-68, 2007.