| Popis: |
In recent years, fuzzy utility mining has become an area of interest due to advancement of human reasoning. With regards to real applications, transactions in a database often involve things, such as transaction time, stamp, and much more. It is also noted that not all products in a store are displayed on the shelf, especially the seasonal ones. This paper, therefore, addresses these issues by presenting an effective framework called temporal-based fuzzy utility mining to give more attention to the transaction period of given items according to the concept of fuzzy utility mining. The temporal-based fuzzy utility mining proposed here is, however, a more complex approach when compared with the traditional fuzzy utility mining. A more complicated model for non-lost upper-bound fuzzy utility is thus proposed for effective mining. Furthermore, based on this model, a two-phase algorithm is developed for temporal-based fuzzy utility mining. Finally, the difference of fuzzy utility item sets with and without consideration of the lifetime of the items is shown by the experimental results under various experimental conditions. |
