Analyse Comparative des Algorithmes de Haute utilité pour représentation concise dans les bases de données transactionnelles

Abstract

High utility item set extraction (HUIM) is an important problem in data mining, consisting of combinations of items that significantly impact a specific metric such as sales or profits. Faced with the exponential growth of data in the world of Big Data, it becomes imperative to design efficient algorithms to extract these sets of high-utility articles quickly and efficiently. In this comparative study, we examined the performance of HUIM algorithms for concise representation. Several large datasets were used to conduct experiments to evaluate the performance of MinFHM, CHUI-MinerMax, EFIM_Closed and CHUD concerning execution time, memory consumption, and number of high utility items mined. According to the results, these algorithms were proven to be able to efficiently extract high utility item sets, with remarkable performance in terms of speed and optimal memory usage.