DWS at the 2016 open knowledge extraction challenge: A Hearst-like pattern-based approach to hypernym extraction and class induction
| Autor: | Simone Paolo Ponzetto, Stefano Faralli |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2016 |
| Předmět: |
Vocabulary
Class induction Hearst patterns Hypernym extraction Linked Open Data Computer science media_common.quotation_subject 02 engineering and technology Representation (arts) Ontology (information science) computer.software_genre 020204 information systems 0202 electrical engineering electronic engineering information engineering Semantic memory media_common business.industry Structured content Linked data Class (biology) Knowledge base 020201 artificial intelligence & image processing Artificial intelligence business computer Natural language processing |
| Zdroj: | Semantic Web Challenges ISBN: 9783319465647 SemWebEval@ESWC |
| Popis: | In this paper we present a system for the 2016 edition of the Open Knowledge Extraction (OKE) Challenge. The OKE challenge promotes research in automatic extraction of structured content from textual data and its representation and publication as Linked Data. The proposed system addresses the second task of the challenge, namely “Class Induction and entity typing for Vocabulary and Knowledge Base enrichment” and combines state-of-the-art lexically-based Natural Language Processing (NLP) techniques with lexical and semantic knowledge bases to first extract hypernyms from definitional sentences and second select the most suitable class of the extracted hypernyms from those available in the DOLCE foundational ontology. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|