Title :
FCA-BASED RULE GENERATOR, a framework for the genetic generation of fuzzy classification systems using formal concept analysis
Author :
Marcos E. Cintra;Maria C. Monard;Heloisa A. Camargo
Author_Institution :
Natural and Exact Sciences Department, Federal Territorial Uni. of the Semi-Arid, Brazil
Abstract :
There is a number of frameworks for the general task of classification available for free usage on the Internet. However, software to generate fuzzy classification systems using the genetic approach is scarce. In this work, we present the FCABASED RULE GENERATOR framework to automatically generate fuzzy classification systems based on a genetic rule selection process. Such rules are extracted from data using a formal concept analysis approach. The FCA-BASED RULE GENERATOR framework includes modules for dataset preprocessing, automatic definition of fuzzy data bases from data, dataset optimization, a module based on formal concept analysis for rule extraction, a genetic algorithm module, as well as a rule base optimization module. The software is described and an example of use is presented.
Keywords :
"Generators","Genetic algorithms","Genetics","Computational modeling","Artificial intelligence"
Conference_Titel :
Fuzzy Systems (FUZZ-IEEE), 2015 IEEE International Conference on
DOI :
10.1109/FUZZ-IEEE.2015.7337950