Automatic Summarization Method of Technical Literature Based On Domain Ontonogy

Abstract

In the field of automatic summarization technology, it is hot but difficult to make full use of the semantic information of text. Owing to the semantic reasoning of ontology, we proposed an automatic summarization method based on domain ontology. The domain ontology is constructed by using knowledge graph as the initial template and the FP-Growth association rule to mine the relations between conceptions. After mapping sentences to RDF triples, the final triples will be given by semantic reasoning based on domain ontology. The summary of technical literature based on domain ontology express the themes more accurately. Taking ROUGE as the evaluation, the method we give is better than the traditional auto-summarization method in average, the ROUGE value of multiple documents summarization is 15% higher than single document summarization.