Lexicon-Based Sentiment Analysis Using Inset Dictionary: A Systematic Literature Review

Abstract

Sentiment analysis is one of the most exciting topics to research in text mining. Lexicon-based sentiment analysis produces good results across a wide range of conversational topics, is easily improved using various sources of information, and does not require additional training. However, only a few review papers have addressed sentiment analysis with specific lexicon dictionaries. As a result, neither data sources nor pre-processing techniques specific to a lexicon dictionary are discussed further. A Systematic Literature Review was conducted to provide a comprehensive and structured lexicon-based sentiment analysis using the Inset dictionary. The literature review resulted in selecting seventeen papers for a detailed study. The findings show that in the last five years, most sentiment analysis research using Inset focused on the health domain. Twitter provides the majority of the data for sentiment analysis. Stopword removal, tokenization, case folding, and stemming are common pre-processing techniques. This study also contained some additional observations from completed research.