The following repository is a first project on sentiment analysis problem by using sentiment140 dataset from Kaggle, containing 1,600,000 tweets extracted using the twitter api (https://www.kaggle.com/datasets/kazanova/sentiment140)
We focus on the resolution of a sentiment analysis problem in two ways: One by using a LSTM model lstm_application.py. And another one by using spacy pipeline spacy_application.py.
- keras == 2.11.0
- tensorflow == 2.11.0
- numpy == 1.23.1
- spacy == 3.5.0
- scikit-learn == 1.1.2
Some specific requirements are needed. First, the file data.csv has to be unzip from the data.zip in the data folder. Secondly, you have to download the trained pipeline en_core_web_sm by writing in the terminal:
python -m spacy download en_core_web_sm