Publication
AI4I 2019
Conference paper

Supporting soccer analytics through hyperknowledge specifications

View publication

Abstract

An issue in semantic search is to infer contextual meaning from terms used in the search query. This is related to appropriately structuring and representing knowledge in a particular domain. In the case of soccer in the entertainment industry, imprecise resolutions when modeling the domain can introduce heavy processing and storage requirements due to the high volume of concepts and content replication demands (e.g., replays). In this work we propose a knowledge retrieval system based on a novel method for modeling the soccer domain: considering contextual meanings to isolate knowledge instances in containers, aiming to reduce the scope of semantic searches by avoiding content replication. It was done by exploiting a hybrid knowledge representation called hyperknowledge that allows the definition of spatiotemporal anchors, and by introducing a query language to ease mapping the contextual meaning into containers stored in hyperknowledge bases.

Date

01 Sep 2019

Publication

AI4I 2019

Authors

Share