← Back to portfolio
Sentiment Analysis System
Sentiment classifier for customer reviews, benchmarking three supervised models.
Problem
Customer reviews carry signal that is expensive to read one by one. Teams need an automated way to gauge sentiment at volume.
Approach
Built an NLP preprocessing pipeline (tokenization, TF-IDF, stemming) and trained supervised models, benchmarking Naive Bayes, SVM, and Logistic Regression against each other with comprehensive evaluation metrics.
Results
- Compared Naive Bayes, SVM, and Logistic Regression head-to-head
- Evaluated with a full metrics suite rather than accuracy alone
Want the full picture, or think this kind of work fits a role you're hiring for?