An extension of the Genetic Iterative Approach for learning rule subsets
Yoel Caises, Enrique Leyva, Antonio González, Raúl Pérez · 2010
Learning fuzzy rules using genetic algorithms has proven to be a feasible way to learn from data with a high level of uncertainly. Some researches in this area are based on the Genetic Iterative Approach, where a genetic algorithm is the main element of an iterative covering scheme, learning one rule in each iteration. The goal of this work is to extend the Genetic Iterative Approach to increase the number of rules extracted in each iteration, as a way to decrease the time for learning. Our proposal is implemented over a fuzzy rule-based algorithm based on the classical Genetic Iterative Approach. This version is also compared with some well-known fuzzy rule-based algorithms.