Student project guidance / Project catalogue
Aspect-Based Sentiment Analysis on E-Commerce Reviews
Extract product aspects and classify sentiment toward each aspect instead of assigning one label to an entire review. Explore CS619 scope and technical guidance.
Domain: Advanced NLP · Technologies: Python, Transformers, PyTorch/TensorFlow, spaCy, Flask/Streamlit
Listed: November 16, 2026 · Reference: Fall 2026 supplied catalogue
Portfolio support fit: Conditional match. Read the feasibility checks below before requesting support.
What this project covers
Extract product aspects and classify sentiment toward each aspect instead of assigning one label to an entire review.
Why this matches Imran’s experience
Support fit: Conditional match. This assessment is based on existing text emotion recognition work. See the related portfolio evidence. This is a proposed support topic, not a claim that this exact project has already been delivered.
Main modules to plan
- Establish an aspect-labelled dataset and baseline.
- Implement aspect extraction and compare sentiment models.
- Build CSV review upload and aspect-level charts.
Technology and feasibility checks
Transformer fine-tuning and sequence labelling go beyond the demonstrated baseline. Confirm dataset access through the student’s VU account and available compute before accepting the full scope.
Testing and demonstration evidence
Evaluate extraction and sentiment separately. Include reviews that praise one aspect while criticizing another and report per-aspect errors.
How to start your SRS and prototype
Bring your current university brief, approved technology choices, deadline and the work you have already attempted. Map each requirement to an actor, stored data, an acceptance criterion and a test. Start with one complete workflow before adding the remaining modules. You can use the CS619 project-phase guidance to organize requirements and the student help desk to discuss a specific blocker.
Catalogue source and availability
This independent scope summary is based on the supplied AllProjectsF26.pdf list. Confirm current availability, detailed requirements and permitted tools with your supervisor or VULMS. The catalogue is not university approval or a booking confirmation.
Support for SRS, design, prototype and final delivery
Discuss requirements and acceptance criteria, architecture and UML, implementation troubleshooting, test evidence and demonstration preparation. Share your university brief, current phase and the specific help you need. Support is independent mentoring; you develop and submit your own assessed work.
Discuss this project on WhatsAppRead the four technical project phases · Share your project details