Our goal is to develop a sentiment classifier using a bidirectional stacked RNN with LSTM/GRU cells for twitter sentiment analysis, from this dataset, which was cleaned in this .
We implemented a class, called LSTM_GRU, where with help of her we managed to experiment with:
- Number of stacked RNNs
- Number of hidden layers
- Type of cells
- Gradient clipping
- Dropout probability
During our experimental procedure, we utilize the Adam optimizer and the Binary Cross-Entropy (BCE) loss function. Each experiment, which took place, was evaluated from learning curves, classification report (which includes precision, recall and F1 score for each class) and ROC Curve's plot. In this last checkpoint of this , we decided to keep LSTM model, which aimed to have a pretty descent performance!
Note that notebook is well reported and was implemented with Machine Learning Library Pytorch. Running's procedure took place on Google Colab, enhanced with Cuda GPU!