Incremental-Eclat Model: An Implementation via Benchmark Case Study

Abstract

Association Rule Mining (ARM) is one of the most prominent areas in detecting pattern analysis especially for crucial business decision making. With the aims to extract interesting correlations, frequent patterns, association or casual structures among set of items in the transaction databases or other data repositories, the end product of association rule mining is the analysis of pattern that could be a major contributor especially in managerial decision making. Most of previous frequent mining techniques are dealing with horizontal format of their data repositories. However, the current and emerging trend exists where some of the research works are focusing on dealing with vertical data format and the rule mining results are quite promising. One example of vertical rule mining technique is called Eclat which is the abbreviation of Equivalence Class Transformation. In response to the promising results of the vertical format and mining in a higher volume of data, in this study we propose a new model called an Incremental-Eclat adopting via relational database management system, MySQL (My Structured Query Language) that serves as our association rule mining database engine in testing benchmark Frequent Itemset Mining (FIMI) datasets from online repository. The experimental results of our proposed model outperform the traditional Eclat with certain order of magnitude.