Incremental maintenance of discovered fuzzy association rules

Fuzzy association rules (FARs) are a recognized model to study existing relations among data, commonly stored in data repositories. In real-world applications, transactions are continuously processed with upcoming new data, rendering the discovered rules information inexact or obsolete in a short time. Incremental mining methods arise to avoid re-runs of those algorithms from scratch by re-using information that is systematically maintained. These methods are useful for extracting knowledge in dynamic environments. However, executing the algorithms only to maintain previously discovered information creates inefficiencies in real-time decision support systems. In this paper, two active algorithms are proposed for incremental maintenance of previously discovered FARs, inspired by efficient methods for change computation. The application of a generic form of measures in these algorithms allows the maintenance of a wide number of metrics simultaneously. We also propose to compute data operations in real-time, in order to create a reduced relevant instance set. The algorithms presented do not discover new knowledge; they are just created to efficiently maintain valuable information previously extracted, ready for decision making. Experimental results on education data and repository data sets show that our methods achieve a good performance. In fact, they can significantly improve traditional mining, incremental mining, and a naïve approach.