Aspect-Based Sentiment Analysis
Principal Investigator: Nisansa de Silva
We propose to advance aspect-based sentiment analysis (ABSA) by developing enhanced models, datasets, and evaluation methods that improve the accuracy and applicability of sentiment classification across multiple domains.
Aspect-Based Sentiment Analysis (ABSA) is a critical task in natural language processing that goes beyond document-level sentiment classification by identifying sentiments tied to specific aspects or attributes of an entity. This project focuses on advancing ABSA through a combination of methodological innovation, dataset creation, and comparative evaluation.
We investigate hybrid transformer-based approaches, such as Instruct-DeBERTa, which integrate instruction tuning and category extraction for improved fine-grained sentiment prediction. Our research also compares traditional rule-based techniques, deep neural networks, and hybrid frameworks to highlight the trade-offs between interpretability and accuracy. Datasets and benchmarks are created to support reproducible experiments and foster wider adoption of ABSA techniques.
Beyond theoretical contributions, we apply ABSA models to practical scenarios such as analyzing product reviews and mobile application feedback, demonstrating their real-world utility for decision-making. By integrating innovative modeling strategies with domain-relevant applications, this work contributes to the broader goal of making sentiment analysis more accurate, interpretable, and adaptable across domains.
Objectives:
- Develop and evaluate deep learning and hybrid models for aspect-based sentiment analysis across diverse textual domains.
- Enhance ABSA frameworks with integrated category extraction to improve fine-grained sentiment detection.
- Compare traditional, rule-based, and neural approaches to identify their strengths and limitations in ABSA tasks.
- Build datasets and benchmarks to support experimentation and progress in ABSA research.
- Apply ABSA techniques to practical domains such as product and mobile app reviews to demonstrate real-world utility.
- Investigate instruction-tuned and transformer-based models (e.g., Instruct-DeBERTa) for robust performance in low-resource settings.
Keywords: Natural Language Processing | Machine Learning / Deep Learning | Law | Education | Aspect-based Sentiment Analysis | Legal Domain | Legal Information Extraction | Information Extraction | Sentiment Analysis |
Publications
Journal Papers
Dineth Jayakody, A V A Malkith, Koshila Isuranda, Vishal Thenuwara, Nisansa de Silva, Sachintha Rajith Ponnamperuma, G G N Sandamali, and K L K Sudheera, "Instruct-DeBERTa: A Hybrid Approach for Aspect-based Sentiment Analysis on Textual Reviews", ICTer, vol. 18, no. 2, pp. 39-50, 2025. doi: 10.4038/icter.v18i2.7290
Isanka Rajapaksha, Chanika Ruchini Mudalige, Dilini Karunarathna, Nisansa de Silva, Gathika Ratnayaka, and Amal Shehan Perera, "Sigmalaw PBSA-A Deep Learning Approach for Aspect-Based Sentiment Analysis in Legal Opinion Texts", Journal of Data Intelligence, vol. 3, no. 1, pp. 101--115, 2021. doi: 10.26421/JDI3.1-1
Conference Papers
Dineth Jayakody, Koshila Isuranda, A V A Malkith, Nisansa de Silva, Sachintha Rajith Ponnamperuma, G G N Sandamali, K L K Sudheera, and Kashnika Gimhani Sarathchandra, "Enhanced Aspect-Based Sentiment Analysis with Integrated Category Extraction for Instruct-DeBERTa", in Proceedings of the 38th Pacific Asia Conference on Language, Information and Computation, 2024, pp. 665--674.
Dineth Jayakody, Koshila Isuranda, A V A Malkith, Nisansa de Silva, Sachintha Rajith Ponnamperuma, G G N Sandamali, and K L K Sudheera, "Aspect-Based Sentiment Analysis Techniques: A Comparative Study", in 2024 Moratuwa Engineering Research Conference (MERCon), 2024, pp. 205--210. doi: 10.1109/MERCon63886.2024.10688631
Sadeep Gunathilaka and Nisansa De Silva, "Aspect-based Sentiment Analysis on Mobile Application Reviews", in 2022 22nd International Conference on Advances in ICT for Emerging Regions (ICTer), 2022, pp. 183--188. doi: 10.1109/ICTer58063.2022.10024070
Isanka Rajapaksha, Chanika Ruchini Mudalige, Dilini Karunarathna, Nisansa de Silva, Amal Shehan Perera, and Gathika Ratnayaka, "Sigmalaw PBSA-A Deep Learning Model for Aspect-Based Sentiment Analysis for the Legal Domain", in International Conference on Database and Expert Systems Applications, 2021, pp. 125--137. doi: 10.1007/978-3-030-86472-9_12
Isanka Rajapaksha, Chanika Ruchini Mudalige, Dilini Karunarathna, Nisansa de Silva, Gathika Rathnayaka, and Amal Shehan Perera, "Rule-Based Approach for Party-Based Sentiment Analysis in Legal Opinion Texts", in 2020 20th International Conference on Advances in ICT for Emerging Regions (ICTer), 2020, pp. 284--285. doi: 10.1109/ICTer51097.2020.9325435
Chanika Ruchini Mudalige, Dilini Karunarathna, Isanka Rajapaksha, Nisansa de Silva, Gathika Ratnayaka, Amal Shehan Perera, and Ramesh Pathirana, "SigmaLaw-ABSA: Dataset for Aspect-Based Sentiment Analysis in Legal Opinion Texts", in 2020 IEEE 15th International Conference on Industrial and Information Systems (ICIIS), 2020, pp. 488--493. doi: 10.1109/ICIIS51140.2020.9342650
Sachintha Rajith Ponnamperuma, Andun S. L. Gunawardana, Thilini Shanika, Prabahth Suminda Pathirana, Shahani Markus, and N. H. Nisansa Dilushan de Silva, "Novel Approach for Perception Analysis In a Learning Environment", in Teaching, Assessment and Learning (TALE), 2014 International Conference on, December. 2014, pp. 148--154. doi: 10.1109/TALE.2014.7062607
Workshop Papers
Dineth Jayakody, Koshila Isuranda, A V A Malkith, Nisansa de Silva, G G N Sandamali, K L K Sudheera, and Sachintha Rajith, "Instruct-DeBERTa: A Hybrid Approach for Enhanced Aspect-Based Sentiment Analysis with Category Extraction", in Eighth Widening NLP Workshop (WiNLP 2024) Phase II, 2024.
Team
External Collaborators: | G G N Sandamali | K L K Sudheera | Shehan Perera | Shahani Markus |
