Problem Mining is a highly useful collection problems of exploiting expand common practice attracted many scientists interested in research. In market research, profit should increasingly interested in exploiting highly useful practice is increasingly focused. Therefore, many mining algorithms useful set high are proposed to improve the efficiency of extraction Huis, in which Efim is the latest algorithm applies multiple techniques to improve the speed and search space as projection transactions, transaction mix, ... However, Efim still costly scans the data stream to determine the candidate is related to consider reducing the efficiency of the algorithm, especially for body database (DB) said. This paper proposed algorithm IEFIM lighting solutions use reverse P-set to reduce the number of transactions should be considered and therefore, reduces the time to explore HUI. Experimental results show that the algorithm significantly IEFIM need to consider the number of transactions and the execution time significantly reduced on sparse databases.
đang được dịch, vui lòng đợi..