This repository contains our codebase for the project "Ideology based on Vocal Patterns" conducted by Yassine Kadiri, Thomas Leble, and Zsolt Pajor-Gyulai under the advice of Prof. Daniel L. Chen and Prof. Elliot Ash for the course DS-GA 1003: Machine Learning and Computational Statistics at New York University.
Our living breathing Google Doc is here.
After submitting our work for the Machine Learning Class, Yassine Kadiri pursued the work over the summer with Professors Ash and Chen in order to produce publishable work.
At first, we had chosen a limited set of words to train our models, in order to have more interpretable results but over the summer we made the choice to train our models over all the words from our dataset. This work can be found in data_gathering_scripts/gathered_data/Regression_all_words.ipynb. It's also summed up in Tone_of_Voice_predicts_political_attitude.pdf