GitHub - arnodunstatter/COSC4337-project: sentiment analysis on financial news headlines

arnodunstatter / COSC4337-project Public

Notifications You must be signed in to change notification settings
Fork 0
Star 0

sentiment analysis on financial news headlines

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 31 Commits
.gitignore		.gitignore
4337 - 1st Deliverable_ Data Pre-processing.pdf		4337 - 1st Deliverable_ Data Pre-processing.pdf
4337 - 2nd Deliverable_ Data Modeling.pdf		4337 - 2nd Deliverable_ Data Modeling.pdf
4337 - 3rd Deliverable_ Data Presentation and Visualization.pdf		4337 - 3rd Deliverable_ Data Presentation and Visualization.pdf
Financial_Sentiment_Analysis_Project.ipynb		Financial_Sentiment_Analysis_Project.ipynb
README.txt		README.txt
ResultingModelAccuracies.csv		ResultingModelAccuracies.csv
generateVisualization.ipynb		generateVisualization.ipynb
models_info.zip		models_info.zip
models_info_with_AF.zip		models_info_with_AF.zip
test.py		test.py

Repository files navigation

Here are some installion commands we used, apart from the usual (pandas, numpy, etc):
pip install wordcloud
pip install nltk
pip install spacy
python -m spacy download en_core_web_sm

It should be noted that the finBERT model is rather RAM hungry.
Particularly the codeblock within the "Now to test over the entirety of the dataset, but in small batches"
header. This section worked when run on a desktop with 6 GB of available RAM, but failed on a laptop with less 
available RAM.