Machine Learning for Computational Social Science

Undergraduate / Graduate

This is a core-course for the MACSS program and it requires Python programming experience (for non-MACSS students, please email the instructor for consultation). This course will teach fundamental skills of applying statistical machine learning models in computational social science tasks. It will focus on understanding the strengths and weaknesses of modern machine learning algorithms as well as their applications in real-world tasks. Topics will include the key techniques in standard machine learning pipelines: data processing (e.g., data representation, feature selection), classification models (e.g., decision trees, logistic regression, naive bayes), regression models (e.g., linear regression), model evaluation (e.g., cross-validation, confusion matrix, precision, recall, and f1 for classification models; RMSE and Pearson correlation for regression models), and error analysis (e.g., data imbalance, bias-variance tradeoff, interpret model performance).