To analyze movie's review and its sentiment and do classification of new review either it is positive or negative using deep learning and LSTM
- Model training - Deep learning
- Method: LSTM
- Visualization toolkit: Tensorboard
In this analysis, dataset used from https://raw.githubusercontent.com/Ankit152/IMDB-sentiment-analysis/master/IMDB-Dataset.csv.
It has 50,000 dataset with 2 columns, Review and Sentiment Review is in a series of string written by movie reviewer Sentiment, either positive or negative is the sentiment set by a human based on the movie review
In this analysis, the visualizing metrics such as loss and accuracy will be appear and keep track on Tensorboard
To open tensorboard,
- Open command prompt
- Activate environment (example: conda activate tf_env)
- Run tensorboard --logdir "PATH of log file"
- If no error, you may open http://localhost:6006/ on browser
Loss & accuracy scalar:
Workflow graph:
- Clone this repository
- Run python file --> deploy.py
- Console will prompt user to input new review
- Press 'Enter'
- Result of sentiment will auto appear below inserted review
Positive sentiment review:
Negative sentiment review:
Enjoy!