Inhaltszusammenfassung

In the last years, the term ontology became more and more popular in computer science. Especially in the semantic web and language based applications, ontologies are used to improve the understanding of freely used language, so user queries and statements can be processed with higher accuracy. Each processed document, therefore, must provide ontology information that must be carefully created by hand for every single domain. In the past, several approaches were introduced to generate ontologi...In the last years, the term ontology became more and more popular in computer science. Especially in the semantic web and language based applications, ontologies are used to improve the understanding of freely used language, so user queries and statements can be processed with higher accuracy. Each processed document, therefore, must provide ontology information that must be carefully created by hand for every single domain. In the past, several approaches were introduced to generate ontologies automatically from unstructured text. Hearst introduced the automatic acquisition of hyponyms by using so-called Hearst Patterns. The so extracted hyponyms and hypernyms can be used for creating ontological hierarchies. This paper focusses on improving the accuracy of the Hearst Patterns by using crowdsourcing platforms and machine learning (ML).» weiterlesen» einklappen

Autoren

Klassifikation

DFG Fachgebiet:
Informatik

DDC Sachgruppe:
Informatik