Efficiency Analysis of ASP Encodings for Sequential Pattern Mining Tasks

Abstract

This article presents the use of Answer Set Programming (ASP) to mine sequential patterns. ASP is a high-level declarative logic programming paradigm for high level encoding combinatorial and optimization problem solving as well as knowledge representation and reasoning. Thus, ASP is a good candidate for implementing pattern mining with background knowledge, which has been a data mining issue for a long time. We propose encodings of the classical sequential pattern mining tasks within two representations of embeddings (fill-gaps vs skip-gaps) and for various kinds of patterns: frequent, constrained and condensed. We compare the computational performance of these encodings with each other to get a good insight into the efficiency of ASP encodings. The results show that the fill-gaps strategy is better on real problems due to lower memory consumption. Finally, compared to a constraint programming approach (CPSM), another declarative programming paradigm, our proposal showed comparable performance.

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

“…The possibility of coupling the power of a high-level declaration with an optimised solver allows users to specify how patterns should be and not how they should be computed. There are different existing approaches that blend together the expressiveness and readiness to use of a declarative system within the problem of pattern mining (Järvisalo 2011;Gebser et al 2016;Samet et al 2017;Paramonov et al 2019;Guyet et al 2014;2016). The main objective of this section is to show the growing interest in declarative-based approaches for pattern mining as an alternative to dedicated algorithms, especially when the main focus is to add expressiveness to the standard problem at hand.…”

“…The first and most common studied constraint is the frequency threshold, where a pattern is validated only if it appears sufficiently often (Agarwal and Srikant 1994). Frequent pattern mining is a fairly well studied problem and effective solutions are available also in the logic programming arena (Järvisalo 2011;Gebser et al 2016;Guyet et al 2014).…”

“…To solve the pattern mining problem introduced above, a guess-and-check resolution scheme, with the support of rule-based languages such as Answer Set Programming (ASP), seems to be quite intuitive and effective. In particular, we have seen that solutions for frequent pattern mining are already available in ASP (Järvisalo 2011;Gebser et al 2016;Guyet et al 2014); however, the constraints we are concerned with in this work to validate patterns come in the form of potentially complex functions, which are difficult, or even impossible, to express in ASP. In order to solve this issue, we resort to recent extensions of ASP systems, such as DLVHEX (Eiter et al 2016;, WASP (Dodaro and Ricca 2020), clingo (Gebser et al 2019), etc., which allow external calculation functions, usually written in Python, to be integrated into ASP programs.…”

Towards Generating Explanations for ASP-Based Link Analysis using Declarative Program Transformations

“…Clingo 3 itself combines a powerful grounder (Gringo) with Clasp (for solving) into an integrated system. For ease of use, and due to its efficiency (e. g., [14,15]), we utilized Clingo in the context of this paper. We assume that the reader has some background knowledge about ASP.…”

Applying Answer Set Programming for Knowledge-Based Link Prediction on Social Interaction Networks

“…Clingo 1 itself combines a powerful grounder (Gringo) with Clasp (for solving) into an integrated system. For ease of use, and due to its efficiency (e. g., Guyet et al, 2018 ; Schäpers et al, 2018 ), we utilized Clingo in the context of this paper.…”