A Modular Framework for Cost-Efficient Aspect-Based Sentiment Analysis Using Small Language Models

Authors

Senthil Kumar, Southern Methodist UniversityFollow
Nibhrat Lohia, Southern Methodist UniversityFollow

Abstract

Aspect-based sentiment analysis (ABSA) links opinions in text to specific product attributes (for example, battery life, screen quality, or delivery speed) rather than only assigning an overall star rating. This level of detail is important in domains such as e-commerce, where teams need to know which features customers praised and which they criticized. Traditional ABSA pipelines have relied on large language models (LLMs), which achieved high quality but were expensive to run and difficult to scale. This study evaluated whether small language models (SLMs) in the 1–3 billion parameter range could serve as a lower-cost alternative. We implemented a modular pipeline in which specialized SLM-based components extracted product aspects from customer reviews and then scored sentiment toward each aspect. We applied this pipeline to Amazon electronics reviews and measured output coverage, sentiment accuracy, latency, and approximate cost. The system produced structured aspect–sentiment pairs for most reviews and achieved a mean absolute error of approximately 0.54 when predicting review-level sentiment compared to user star ratings. It also ran on rented GPU endpoints at an estimated cost of about $0.12 per 100 reviews. These results provided initial evidence that coordinated SLMs can deliver actionable feature-level sentiment signals at materially lower cost than a single large model.

Recommended Citation

Kumar, Senthil and Lohia, Nibhrat (2025) "A Modular Framework for Cost-Efficient Aspect-Based Sentiment Analysis Using Small Language Models," SMU Data Science Review: Vol. 9: No. 3, Article 4.
Available at: https://scholar.smu.edu/datasciencereview/vol9/iss3/4