Learning approximate fuzzy rules from training examples
Tzung‐Pei Hong, Tzu-Ting Wang, Been-Chian Chien · 2002
Ziarko (1993) proposed the variable precision rough set model to deal with noisy data. This model identifies the relationships among data using crisp attribute values. However, data with quantitative values are commonly seen in real-world applications. In this paper, we propose a new algorithm to produce a set of maximally general fuzzy rules for an approximate coverage of training examples based on the variable rough set model from noisy quantitative training data. The proposed algorithm first transforms each quantitative value into a fuzzy set of linguistic terms using membership functions and then calculates the fuzzy /spl beta/-lower and the fuzzy /spl beta/-upper approximations. The maximally general fuzzy rules for approximately covering the given data are then generated based on these fuzzy approximations by an iterative induction process.