Abstract Existing utility mining algorithms consider only frequency, productivity and quantity of every item purchased for utility calculation. No attention has been given to the length of transactions. Uncovering high average utility itemsets from the transaction dataset solves the above issue. In some circumstances, things may appear in reality with a negative unit value. For example, a retail store may trade items at a loss to energize the trade of similar products. To resolve the issue, we propose an effective mechanism named appropriate average high utility itemset mining with negative utilities (AAHUIM-NU) which creates high quality decision makers. To solve the issue of data set repository and numerous sweeps of the data set, proposed algorithm involves number of lists for caching the relevant data which possesses only less storage. It utilizes a minimized transaction utility, decreased maximal utility, and generalized tight upper bounds to find out the pruning threshold and to minimize the running time and memory. Exploratory outcomes on datasets demonstrate that AAHUIM-NU is productive in terms of processing time, memory usage, and versatility. It performs well on thick datasets which has an excessive number of long transactions. The experimental results are recorded in tables and given in this paper.
Alan : Mühendislik
Dergi Türü : Uluslararası
Benzer Makaleler | Yazar | # |
---|
Makale | Yazar | # |
---|