Portal de Periódicos da CAPES

Deriving Pertinent Knowledge through Sentiment Analysis and Linking with Relevant Documents

2021; Sarada Ranganathan Endowment for Library Science (SRELS); Linguagem: Inglês

10.17821/srels/2021/v58i5/160674

0976-2477

Apala Chatterjee, Shampa Mahato, Sunil Kumar Chatterjee,

Topic Modeling

Purpose: This study aims to explore pertinent knowledge through the Sentiment Analysis technique and to link with relevant, pin-pointed documents. Design/Methodology/Approach: While information is essential ‘information overload’ is a big problem when we search for specific information. To get rid of psychological stress, mistakes in decision making or disregarding of relevant information, a methodology has been developed which may be suitable for researchers to extract pertinent knowledge from huge amount of research publications in a particular domain (‘climatology’ has been chosen for demonstration) within the shortest possible time. The study presents, how exactly relevant information can be retrieved there through sentiment analysis and through which a preliminary knowledge base can be gained. For this, ‘R’ software has been used to do the desired manipulation on the collected data. The steps involve pre-processing of introductory text, tokenization, polarity detection and analysis of text through sentiment analysis. Findings: It has been found that knowledge derived through sentiment analysis and abstract of the linked documents fairly match with each other, which validates the relevance and importance of the linked documents. Again, the impact factor of the prestigious journal having global coverage, where most of the linked documents were published also shows the importance of the linked documents/papers.