Battle of Sentiment Lexicons: Wordnet, Sentiwordnet, Textblob and Vader in Web Forum Analysis
Keywords:
Text Analysis, Sentiment Analysis, Housing, Lexicon, NLP, Decision MakingAbstract
Introduction: Sentiment research is essential for evaluating public opinion in a number of industries, including the high-rise real estate sector. Comprehending public opinion offers crucial perspectives for interested parties, facilitating data-driven and knowledgeable decision-making. The usefulness and efficacy of lexicon-based sentiment analysis methods in gauging public opinion against high-rise buildings in Malaysia are investigated in this research.
Objectives:The study's objectives are to examine the efficacy of four lexicon-based sentiment analysis tools: Valence Aware Dictionary and Sentiment Reasoner (VADER), WordNet, TextBlob, and SentiWordNet. In recognizing positive, negative, and neutral attitudes stated in public reviews, it aims to assess their categorization performance and accuracy.
Methods: To examine public opinion on high-rise buildings, the study used a real-world case study. The four lexical dictionaries were used to categorize reviews, and their performance was evaluated by contrasting their capacity to recognize neutral, negative, and positive attitudes.
Findings: The investigation showed that a greater percentage of positive reviews than negative or neutral ones were detected by all four lexicon-based techniques. Positive feelings were most prevalent in WordNet, closely followed by SentiWordNet and VADER. The most successful method for detecting negative attitudes was VADER, which was followed by WordNet and SentiWordNet. VADER found the most evaluations for neutral feelings, but SentiWordNet and WordNet found less. When compared to the other tools, VADER performed better overall. The research did, however, also draw attention to VADER's performance shortcomings, including limited vocabulary, trouble identifying sarcasm, ambiguity, and difficulties with misspelled or short phrases, which sometimes resulted in misclassifications.
Conclusions: For scholars, practitioners, and politicians examining public opinion about high-rise real estate in Malaysia, this study offers insightful information. Through more accurate sentiment categorization, stakeholders may improve their decision-making processes by comprehending the advantages and disadvantages of lexicon-based sentiment analysis technologies.