A Graph based Approach for Keyword Extraction from Documents

This paper presents an unsupervised approach for keyword extraction from a document. Summarizing a research paper is not an easy task, it is time-consuming. For making summarization of documents much easier keywords are used. Keyword helps in summarizing a large amount of textual data. Keyword extraction is a process in which a set of words are selected that gives the context of the whole document. Here, we are proposing a graph-based model approach for extracting keywords from a research paper. The proposing method is a combination of Rapid Automatic Keyword Extraction (RAKE) algorithm and Keyword Extraction using Collective Node Weight (KECNW), in which candidate keywords are extracted using RAKE algorithm and selection of keywords using KECNW model. The final output of this methods are keywords extracted from a document with their rank.