Concept-Based Compound Keyword Extraction Based on Using Sentential Distance, Conceptual Distance and Production Rules: Calculation of the Keyword Importance

摘要

Humans can read a document and conceptually organize its contents into few compound keywords that capture the essence of the topic of a document. Based on this information, this study proposes a method for extracting keywords that gives the gist of a document. It uses a set of academic papers as test data to set up a concept-based production rule for forming compound keywords even when author-provided keywords do not appear in the text body of a document. It also proposes a method of calculating the importance of keyword in order to refrain from extracting meaningless keywords. Also the validity of extracted keywords was tested using a data set of thesis paper tides and summaries in the field of natural language processing and speech recognition. Comparison of the author-provided keywords to the keyword results of the developed system showed that the developed system was very useful with an accuracy rate as good as up to 96 %.