Extended High-Utility Pattern Mining: An Answer Set Programming-Based Framework and Applications

Abstract Detecting sets of relevant patterns from a given dataset is an important challenge in data mining. The relevance pattern, also called utility the literature, subjective measure and can be actually assessed very different points view. Rule-based languages like Answer Set Programming (ASP) seem well suited for specifying user-provided criteria to assess pattern form constraints; moreover, declarativity ASP allows easy switch between several order analyze In this paper, we make steps toward extending notion High-Utility Pattern Mining; particular, introduce new framework that classes not considered previous literature. We show how recent extensions with external functions support fast effective encoding testing framework. To demonstrate potential proposed framework, exploit it as building block definition innovative method predicting ICU admission COVID-19 patients. Finally, extensive experimental activity demonstrates both quantitative qualitative point view effectiveness approach.