Mandolin: A Knowledge Discovery Framework for the Web of Data

Tommaso Soru,Diego Esteves,Edgard Marx,Axel-Cyrille Ngonga Ngomo

DOI: https://doi.org/10.48550/arXiv.1711.01283

2017-11-03

Databases

Abstract:Markov Logic Networks join probabilistic modeling with first-order logic and have been shown to integrate well with the Semantic Web foundations. While several approaches have been devised to tackle the subproblems of rule mining, grounding, and inference, no comprehensive workflow has been proposed so far. In this paper, we fill this gap by introducing a framework called Mandolin, which implements a workflow for knowledge discovery specifically on RDF datasets. Our framework imports knowledge from referenced graphs, creates similarity relationships among similar literals, and relies on state-of-the-art techniques for rule mining, grounding, and inference computation. We show that our best configuration scales well and achieves at least comparable results with respect to other statistical-relational-learning algorithms on link prediction.