Authors : R. Radha and S.P. Rajagopalan
Abstract: A data mining technique usually generates a large amount of patterns and rules. However, most of these patterns are not interesting from a user`s point of view. Beneficial and interesting rules should be selected among those generated rules. This selection process is what we may call a second level of data mining. To prevent the user from overwhelming with rules, techniques are needed to analyze and rank them based on their degree of interestingness. There are two aspects of rule interestingness, objective and subjective aspects. In this study, we are concentrating on both the subjective and objective measures of interestingness. A generic problem of finding the interesting ones among generated rules is addressed and a new mathematical measure for finding the interestingness is explained. We have used fuzzy linguistic terms for the attributes, so that the semantics of such rules are improved by introducing imprecise terms in both the antecedent and the consequent, as these terms are the most commonly used in human conversation and reasoning. The terms are modeled by means of fuzzy sets defined in the appropriate domains. However, the mining task is performed on the fuzzy data. These fuzzy association rules are more informative than rules relating precise values.
R. Radha and S.P. Rajagopalan , 2007. Assessing the Interestingness of Discovered Knowledge Using a Hybrid Approach Based on Fuzzy Concepts. International Journal of Soft Computing, 2: 243-248.