Improving the Efficiency of a Classifier based on Fuzzy Association Rules using Genetic Rule Selection

Improving the Efficiency of a Classifier based on Fuzzy Association Rules using Genetic Rule Selection

Mojtaba Heravi1 Tabassom Azimi galeh2 Hessam Zandhessami3

1) Master of Science, Knowledge Engineering and Decision Sciences; Member of Ahvaz Electric Power Distribution Company; Member of Young Researchers and Elite Club, Qazvin Branch, Islamic Azad University, Iran. Email:
2) Master of Science, Business Management (Marketing); Member of Ahvaz Electric Power Distribution Company. Email:
3) Assistant Professor, Department of Industrial Management, Qazvin Islamic Azad University. Email:

IOI Code : XYEZ-AKZAD
Publication : International Conference On Research Science And Technology(rstconf.com)
Abstract :
    Classification is one the most popular techniques in real-world applications. Among different classification techniques, extracting Association Rules is very useful and can be applied in different domains. In this method, probabilistic relations are detected using support and confidence parameters. However, one of the main drawbacks is the number of extracted rules which computationally is expensive and leads to imprecise results. In the proposed method, Genetic Algorithms are adapted to more efficiently select the probabilistic rules. The results underscore that proposed CFAR-GRS outperforms CFAR and other similar methods such as CBA، CMAR and CPAR and generates the less number of rules. Interestingly, due to use of an evolutionary algorithm in efficiently selecting the rules, obtained classes are more accurate.

Keywords : classification fuzzy association rules genetic algorithm genetic rule selection optimization