M.C. Monard and J.S. Sichman (Eds.): IBERAMIA-SBIA 2000, LNAI 1952, pp. 33-42, 2000. © Springer-Verlag Berlin Heidelberg 2000 Integrating Rules and Cases in Learning via Case Explanation and Paradigm Shift Alneu de Andrade Lopes and Alípio Jorge LIACC - Laboratório de Inteligência Artificial e Ciências de Computadores Universidade do Porto - R. do Campo Alegre 823, 4150 Porto, Portugal E-mail: {alneu,amjorge}@ncc.up.pt, http://www.ncc.up.pt/liacc/ML Abstract. In this article we discuss in detail two techniques for rule and case integration. Case-based learning is used when the rule language is exhausted. Initially, all the examples are used to induce a set of rules with satisfactory quality. The examples that are not covered by these rules are then handled as cases. The case-based approach used also combines rules and cases internally. Instead of only storing the cases as provided, it has a learning phase where, for each case, it constructs and stores a set of explanations with support and confidence above given thresholds. These explanations have different levels of generality and the maximally specific one corresponds to the case itself. The same case may have different explanations representing different perspectives of the case. Therefore, to classify a new case, it looks for relevant stored explanations applicable to the new case. The different possible views of the case given by the explanations correspond to considering different sets of conditions/features to analyze the case. In other words, they lead to different ways to compute similarity between known cases/explanations and the new case to be classified (as opposed to the commonly used fixed metric). 1 Introduction The integration of rules and cases for classification tasks is a recurrent theme in machine learning (Domingos, 1996; Golding & Rosenbloom 1996). Typically, flat case representations, such as vectors of values, have been used for integration. Currently, the trend is towards multistrategy learning systems with richer representation languages dealing with natural language processing, information retrieval from the Web, etc, (Jorge & Lopes 1999), (Doan et al. 2000). Here we discuss in detail two techniques for rule and case integration, already presented in (Lopes & Jorge, 2000). The first technique consists in using a case based approach when the rule language is clearly exhausted. The second one is by producing a set of explanations for each case, where each explanation is a generalization of the case. These explanations, although syntactically similar to rules, are used to classify new cases in a way similar to usual case based approaches. The contribution of the two techniques are evaluated separately and in combination. The classification task considered is the task of morpho-syntactic disambiguation, usually known as part-of- speech tagging.