There are a lot of reviews we all read today- to hotels, websites, movies, etc. So in order to check the sentiment present in the review, i.e. either the review or the whole set of reviews are good or bad we have created a python project which tells us about the positive or negative sentiment of a review.
We have used UMLfit model for text classification. It consists of 3 LSTM layers and is already trained on more than 100 million words from Wikipedia. The model architecture can be explained in the diagram below.
- Numpy
- Pandas
- fast.ai
- tkinter
- matplotlib
- ULMFit Model https://docs.fast.ai/
- Deep Learning Content - https://github.com/smlra-kjsce
- Tkinter Docs - https://tkdocs.com/
- Matplotlib documentation - https://matplotlib.org/3.2.1/contents.html
- LSTMs- The basics of Natural Language Processing https://www.youtube.com/watch?v=9TFnjJkfqmA