Keyword and key-phrase extraction using a graph-based algorithm

Abstract

The technique of keyword and key-phrase extraction involves feeding a text to a computer, which then recommends words and phrases that are relevant to the text's content of documents. The use of keyword and key-phrase extraction techniques is widespread, especially when it comes to information retrieval. This is particularly interesting since individuals employ keywords and key-phrases to access important information. Numerous methods have been created for English texts. However, very few attempts have been made to extract Bengali keywords and to understand context. Four distinct datasets that were gathered from news portals are used in this study to demonstrate a Graph-based algorithm for keyword and key-phrase extraction. Additionally, we have demonstrated if studying co-occurrences permits documenting the creation of each Bengali text. However, manipulating it takes more time and frequently results in messy visualisations.