As part of the Stanford in Paris study abroad program and in collaboration with the Institut Supérieur d'électronique de Paris, my team member and I created a machine learning model to detect terrorist supporters on Twitter. We achieved an accuracy of 92%.
- leveraged the Google translate API to convert French and Arabic tweets to English
- cleaned and organized the data by converting it to CSVs and tokenizing and stemming it
- used Word2Vec and sentiment analysis
partnered with Eliza Deberry, mentorship from Patricia Conde-Cespedes
Check out our published research: