Keyphrase Extraction as Topic Identification Using Term Frequency and Synonymous Term Grouping

摘要

Keyphrase are usually used as a representative of in the document. This paper presents a method to improve keyphrase extraction by using synonymous term grouping. Topic identification is recognised by term frequency for keyphrase extraction. We utilize a language model including linguistic patterns and language knowledge such as morphology syntax. The language model is a probability of word sequence. The focus unit is a pattern of noun adjective combination The proposed method consist of five processes i.e. preprocessing, candidate selection, semantic-based topic clustering, topic ranking, and keyphrase selection. This experimental result has precision value 54.44 from dataset of IEEE and 39.99 from dataset of SamEval.