Ahmad Taufik, Nursal and Mohd Faizal, Omar and Mohd Nasrun, Mohd Nawi and Mohamad Sukeri, Khalid and Mohd Hanizun, Hanafi and Rafikullah, Deraman (2025) Battle of sentiment lexicons: Wordnet, sentiwordnet, textblob and vader in web forum analysis. Journal of Information Systems Engineering and Management, 10. pp. 84-93. ISSN 2468-4376. (Published)
|
Pdf
Battle of sentiment lexicons_Wordnet, sentiwordnet, textblob and vader in web forum analysis.pdf Available under License Creative Commons Attribution. Download (791kB) | Preview |
Abstract
Sentiment analysis plays a crucial role in assessing public perception across various sectors, including the high-rise property industry. Understanding public sentiment provides essential insights for stakeholders, enabling data-driven and informed decision-making. This study examines the applicability and effectiveness of lexicon-based sentiment analysis tools in measuring public sentiment toward high-rise properties in Malaysia. Objectives: The study aims to investigate the effectiveness of four lexicon-based sentiment analysis tools, namely WordNet, TextBlob, SentiWordNet, and Valence Aware Dictionary and Sentiment Reasoner (VADER). It seeks to evaluate their classification performance and accuracy in identifying positive, negative, and neutral sentiments expressed in public reviews. Methods: The research employed a real-world case study to analyze public sentiment toward high-rise properties. Reviews were classified using the four lexicon dictionaries, and their performance was assessed by comparing their ability to identify positive, negative, and neutral sentiments. Results: The analysis revealed that all four lexicon-based tools identified a higher proportion of positive reviews than negative or neutral ones. WordNet recorded the largest number of positive sentiments, closely followed by SentiWordNet and VADER. When identifying negative sentiments, VADER emerged as the most effective, followed by WordNet and SentiWordNet. For neutral sentiments, VADER detected the highest number of reviews, while SentiWordNet and WordNet identified fewer instances. Overall, VADER demonstrated superior performance compared to the other tools. However, the study also highlighted limitations in VADER’s performance, such as restricted vocabulary, difficulty in detecting sarcasm, ambiguity, and challenges with misspelled or short terms, which occasionally led to misclassifications. Conclusions: This study provides valuable insights for researchers, practitioners, and policymakers involved in analyzing public sentiment toward high-rise properties in Malaysia. By understanding the strengths and limitations of lexicon-based sentiment analysis tools, stakeholders can enhance their decision-making processes through more precise sentiment classification Copyright © 2024 by Author/s and Licensed by JISEM.
| Item Type: | Article |
|---|---|
| Additional Information: | Indexed by Scopus |
| Uncontrolled Keywords: | Decision Making; Housing; Lexicon; NLP; Sentiment Analysis; Text Analysis |
| Subjects: | H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management T Technology > T Technology (General) |
| Faculty/Division: | Faculty of Industrial Management |
| Depositing User: | Mr Muhamad Firdaus Janih@Jaini |
| Date Deposited: | 17 Feb 2025 08:04 |
| Last Modified: | 17 Feb 2025 08:04 |
| URI: | http://umpir.ump.edu.my/id/eprint/43834 |
| Download Statistic: | View Download Statistics |
Actions (login required)
 |
View Item |