This project explores a variety of feature-selection and machine-learning techniques using World Bank population data. We use LASSO L1 regularized least-squares regression to choose a 5-sparse linear combination of countries to predict population for any other given country. Finally we use the results to draw insight into the ability of a small number of large-population countries to embody information which predicts other populations.
Data: World Bak Data