Project
Aspect-Based Sentiment Analysis of Tourist Reviews Using BERT (Undergraduate Thesis)
Built a BERT model that identifies what tourists like or dislike about specific aspects of a place from their reviews.
- Researched and built an aspect-based sentiment analysis model using BERT to understand tourist opinions about specific aspects of a destination (for example: services, accommodation, food, transport).
- Manually created a new travel-related dataset by collecting real tourist reviews, then cleaning, organizing and labeling them by aspect and sentiment.
- Trained and evaluated the model on this dataset and interpreted the results to find out what tourists talk about most and how they feel about it.
- Presented and successfully defended the research at my undergraduate thesis defence.
- Currently working on two research publications expected to come from this work.
- Tools used: Python
- BERT
- [add others you used
- e.g. Hugging Face Transformers
- PyTorch
- Pandas
- Excel]