Mining Frequent Weighted Utility Patterns in Dynamic Quantitative Databases

Mining frequent weighted utility patterns (FWUPs) is an important activity in data mining to discover frequent patterns from quantitative databases with considering the weight (mean the importance) of each item. In practical, the weight of each item in a database may varies with time. For example, the weights of the products in a store may change by each month, each quarter, or each year. Therefore, this paper will concern with that issue. At first, we introduce a new problem of mining FWUPs with the dynamic weighted items in the quantitative databases (called dynamic quantitative databases - dQDB). Then, we propose an algorithm named dFWUT using Tidset data structure to solve this problem. Finally, some experiments were conducted to compare the proposed algorithm with a modification of WUN-Miner (denoted by WUN-Miner-Mod), a state-of-the-art algorithm for mining FWUPs in dQDB, in terms of time and usage memory. The experimental results indicate that the dFWUT algorithm outperforms the WUN-Miner-Mod algorithms for mining FWUPs in dQDB.