Predicting-Political-Revolutions
PublicAn in-depth study of how attributes of political protests are an extremely valuable predictor of government change to come. The project uses Data Science techniques for detailed data wrangling, cleaning, and exploration before applying Machine Learning models. After applying techniques such as encoding, binning, log transformations, addressing class imbalance and hyperparameter tuning, the final model provides an F1 score of 0.80 and an accuracy of 0.96.
data-analysisdata-sciencedatasetmachine-learningpoliticspredictive-modelingpythonrandom-forestsqlxgboost
Heure de création：2021-05-14T11:06:55
Heure de mise à jour：2024-01-14T19:55:10
4
Stars
0
Stars Increase