Fi-Eclat: An enhancement of incremental Eclat algorithm

Nowadays, there is a large amount of data growth at all organizational scales. To find useful information, the right data mining technique needs to be used. One of the popular techniques is called the equivalence class transformation (Eclat) algorithm which is gaining the benefits of the vertical data format used in the association rule mining technique. However, like horizontal data representation, vertical data still suffering from huge memory consumption. The technique then has been extended by using the Incremental Eclat (i-Eclat) algorithm which is embedded with the incremental approach. Inspired by the dynamic data transactions in a database, a new algorithm has been adopted to optimize the performance of the current algorithm called the Fast Incremental (Fi-Eclat) algorithm. The research proposes an optimization of the performance of the i-Eclat algorithm. It relies on the incremental approach implemented on the database that consists of five (5) benchmark datasets from Frequent Itemset Mining (FIMI). The outcome of the experiment indicates an improvement of 32% in time execution for the proposed algorithm.