A project that labels amazon review 'Positive review' or 'Negative review'.
Accurracy 80.39%
Project consisting of:
- AI trained sentiment model using Random Forest classifier on Google colab and exported as .pkl format, and accessible via python API
- API server created on Python FastAPI
- Flutter web built on Flutter framework
- Web server created on Python FastAPI
- Nginx file configuration for two servers to run on 8000 and 8080 ports
Currently hosted on AWS EC2 instance and accessible via link
This is a university coursework for Torrens University ISY503 Intelligent Systems subject.
The projec is not ready for production stage, it's merely a prototype.
Done by Petras.janulevicius@gmail.com
There is a 37 pages report on how everything is working and assembled. Accessible on demand.