Sentiment Analysis of Social Network for Information Professionals: A Case Study of LisLinks Discussion Forum

Abstract

This study aims to analyze members' sentiments in India's one of the most used Library and Information Science forums, the LisLinks discussion forum. We retrieved 10,420 discussion posts and their five replies to analyze the sentiment: Positive, Negative, and Neutral, using the RapidMiner tool through a lexicon-based approach. The study results show that 64% of the replies have neutral sentiment, 32% have positive sentiment, and only 4% have a negative sentiment. The more neutral sentiment indicates that the forum members are not actively participating in the discussion site. When we compare the sentiment of the five replies, the results show an increment in neutral sentiment from the first to the fifth. The undertone of the research finding pertains to the users' gradual decremental use of this discussion platform. This study may aid administrators of the forum in grasping the feelings and opinions of the members to enhance the quality of the discussion forum.

Keywords

• discussion forum
• sentiment analysis
• social network
• lislinks
• rapidminer
• octoparse