Mining frequent weighted utility patterns with dynamic weighted items from quantitative databases

Mining frequent weighted utility patterns with dynamic weighted items from quantitative databases

연구 분야: Databases

학회: Applied Intelligence

The mining of frequent weighted utility patterns (FWUPs) is an important task in the field of data mining that aims to discover frequent patterns from quantitative databases while taking into account the importance or weight of each item. Although there are many approaches that have been proposed to solve this problem, all of these methods focus on databases in which the weight of each item is fixed. In real-life situations, the weight of each item may change over time; for example, the weights of the products in a store may change every month, every quarter, or every year. This is an important aspect that previous studies have not considered. In this paper, we first introduce a new problem that involves mining FWUPs with dynamic weighted items from quantitative databases (called dynamic quantitative databases, dQDBs). Following this, we propose an algorithm called dFWUT that uses a tidset data structure to solve this problem. Next, an algorithm called dFWUNL is developed that uses a new data structure called a WUNList to mine FWUPs from dQDBs. Finally, experiments on multiple databases are carried out to show that the proposed method is more efficient than another state-of-the-art algorithm in terms of running time and memory usage, especially for dense datasets or sparse datasets with a small mining threshold.