Multi-domain training improves sentiment analysis for Marathi texts
Enhancing Zero-Shot Domain Transfer in Low-Resource Sentiment Analysis: A Case Study in Marathi
Computation and LanguageMachine Learning
Summary
The paper looks at how computers can better understand feelings in Marathi text, a language with little data available. Models trained on just one type of text don’t work well on other kinds of text. The authors found that training the models on texts from two different areas helps them understand new topics about 8% better. Adding a third type of text helps a little more, but not as much as the second did. This means using examples from different topics makes these computer programs more reliable for less-studied languages.
What this means in practice
- •For language technology developers: Build more reliable Marathi sentiment models that perform well on new topics using multi-domain training data.
- •For social media monitoring teams: Improve automatic opinion detection in Marathi content across diverse subjects by leveraging models trained on mixed-domain data.
Authors
Tanvi Somani, Aryan Babare, Samyak Bora, Raviraj Joshi
Abstract
Sentiment analysis is a key task in Natural Language Processing that determines the polarity of opinions in text. For low-resource languages like Marathi--where limited data prevents the direct use of standard statistical methods--models are often trained on single-domain datasets. While this yields strong in-domain performance, it leads to poor generalization to unseen domains. Recent studies highlight the need for domain-robust systems that can handle diverse real-world content through multi-domain modeling. This study investigates how multi-domain training improves sentiment classification for unseen Marathi domains. Models are trained using single-domain, two-domain, and three-domain dataset combinations. Results show that two-domain training improves accuracy on unseen domains by about 8\% compared to single-domain training. Adding a third domain provides further improvement, though the gain is smaller, averaging around 2\% for some domains. These results suggest that training with data from multiple domains helps build more robust sentiment analysis models for low-resource languages, though the benefits reduce as more domains are added.