Hybrid Algorithm with Map Reduce Framework to Mine Distributed Association Rules from Big Data

When its big data, it's extremely large data sets. These datasets are analyzed computationally to reveal patterns, trends, and associations. That could be related to human behavior and interactions and product reordering ratio or understanding the symptoms or related signs of a disease. Many data analysis techniques are already defined and used by researchers. Results are still showing the scope of improvement. Based on the volume, variety, and velocity of data, the techniques are needed to be used or improved. Association rule mining is one of the technique to solve issues of accuracy in retrieved results. They are used to detect changes in customer behavior, buying trends and reasons that affect such process. Researches till date has proven the results are better than the earlier one. Though several methods have been suggested for the extraction of association rules, problems arise when data is in growing pattern with large volume. To overcome such issue, we propose, in this paper, a hybrid approach based on ARM techniques with Map Reduce framework, modified for processing large volumes of data in an increasing manner. Furthermore, because real life databases lead to a huge number of rules' including many redundant rules, our algorithm proposes to mine a compact set of rules with no loss of information. The results of experiments tested on large real world datasets highlight the relevance of mined data. Additionally in this research, the experiments are performed in continuous growing data which still yields comparative results.