This classifier can predict if a comment has 5 star ratings or not just by reading comment texts. The data set we used is fairly large with more than 5 million samples. To speed up the computations process, we will use 500,000 samples for training, 10,000 for the dev-test set and 10,000 for the test set.
Software Engineer
Sydney, Australia
Email: hossain.tanvir.m@gmail.com
Originally this project was done by me in June, 2018.