A Hybrid Integrated Model for Big Data Applications Based on Association Rules and Fuzzy Logic: A Review

There is a real increasing in the generating of data from different sources. Data mining (DM) is a useful method to elicit valuable information. Association rule mining (ARM) can assist in finding patterns and trends in big data. Also fuzzy logic plays a main role as assistance technique in handling big data issues. This review paper produces recent literature on hybridization regarding the association rule mining or other DM methods such as classification and clustering and fuzzy logic techniques in big data. Whereas a hybrid model of association rule and fuzzy logic is suggested to get a valuable knowledge for big data applications at good accuracy and less time, with the aid of distributed framework for big data handling (Hadoop, Spark and MapReduce). Different techniques and algorithms were used in these works and evaluated according to accuracy, sensitivity, recall and run time with a various result as Specificity = 86%, Sensitivity = 80% and F-measure = 2.5, or achieving high accuracy and shorter runtime compared to other methods and 98.5accuraccy of fitness function in pruning redundant rules. At the end of the paper we present the most used and prominent techniques that assist in providing a useful and valuable knowledge in different domains from a huge, unstructured and even heterogeneous data. The paper will be beneficial to the researches who interesting in the field of mining big data.