— The impressive growth of the Web in the last decade has made it becomes one of the most popular news sources. Web-users nowadays can access an unlimited amount of news items but only a subset of them meets their interest. The existing news systems on Web reveal more and more limitations in information search because news items are presented only for human consumption. This paper presents BKSport, a sports news aggregation system that provides users with the ability to find relevant news articles through queries that are formulated in natural language. Our contribution consists of a system development approach, which is totally based on Semantic Web technologies. First, semantic annotations about news item are created using an ontology and knowledgebase for sports domain. Additionally, we propose a method to transform natural language questions into SPARQL queries to carry out semantic search on these semantic annotations. Last, the system has been positively evaluated with respect to precision, based on a set of pre-defined questions belonging to various categories.
Paper
Full text
Natural Language Questions for Semantic Web-Based News Aggregator
Semantic Scholar · Computer Science · 2019
Abstract
— The impressive growth of the Web in the last decade has made it becomes one of the most popular news sources. Web-users nowadays can access an unlimited amount of news items but only a subset of them meets their interest. The existing news systems on Web reveal more and more limitations in information search because news items are presented only for human consumption. This paper presents BKSport, a sports news aggregation system that provides users with the ability to find relevant news articles through queries that are formulated in natural language. Our contribution consists of a system development approach, which is totally based on Semantic Web technologies. First, semantic annotations about news item are created using an ontology and knowledgebase for sports domain. Additionally, we propose a method to transform natural language questions into SPARQL queries to carry out semantic search on these semantic annotations. Last, the system has been positively evaluated with respect to precision, based on a set of pre-defined questions belonging to various categories.