patrikasvanagas/bi-lstm-glove-sentiment-classification

End-to-end review sentiment classification with text preprocessing, Bidirectional Long Short-Term Memory networks and Glove embeddings.

Bidirectional-LSTM-Sentiment-Classification

This repository provides an example of an end-to-end approach to sentiment classification using Bidirectional Long Short-Term Memory networks with pre-trained Glove embeddings. The workflow involves several preprocessing steps, including expansion of contractions and extraction of sentiment from multiple files, before implementing the Bi-LSTM model for classification. Additionally, experiments with early stopping, test sequences and cross-validation are carried out. The dataset, available in this repository, originates from the following paper:

Ding, X., Liu, B. and Yu, P. S. (2008) ‘A Holistic Lexicon-Based Approach to Opinion Mining’. In Proceedings of the 2008 International Conference on Web Search and Data Mining (WSDM '08). DOI: 10.1145/1341531.1341561.

To run the notebook with access to a free GPU, click on the 'Open In Colab' badge above.