This is a project on a large corpus containing tweets, movie reviews, book reviews and crawled reviews from various shopping sites. Where word embeddings ConceptNet Numberbatch is used. Which can be downloaded from https://github.com/commonsense/conceptnet-numberbatch
- Analysis of tweets, movie reviews, book reviews and crawled reviews from various shopping sites
- Seperate good and bad reviews
- Accuracy of the network rmse analysis is more than 98%
- Efective Embedding and Encoding method
- Efective data pre processing
- Accurate result
- Efficient and fast data rendering
- Efficient memory management
- Concurrent design with CUDA
- Python
- keras
- Tensorflow
- CUDA
- Anaconda
- Download this project
- DOwnload numberbatch-en-17.02.txt from https://github.com/commonsense/conceptnet-numberbatch
- DOwnload data set from https://www.kaggle.com/arshjat/question2/
- To train the model run rnn.py