Fouille Interactive de Motifs à Haute Utilité
Abstract
In recent years, HUI (High Utility Itemset) mining has become a substantial area of research in pattern mining. HUI Mining is a family of methods used to mine patterns by weighting items according to their importance in the transactions in which they occur. HUIs are, therefore, patterns that hold clear importance for the user. The most current approaches require the user to provide a set of utilities linked to both items and transactions in addition to the dataset. Providing this dataset is straightforward. However, obtaining the utilities can be challenging in certain scenarios. In this paper, we introduce a utility-free method for interactively mining high-utility itemsets. Our approach consists of learning the utilities from the user feedback and using the learned utilities to gradually enhance the set of HUIs.