Country Happiness

June 2025

What is it?

The paper embedded below is my final group project for the UChicago computer science course CMSC 35300: Mathematical Foundations of Machine Learning. 

Our objective for this project was to use the machine learning techniques we studied in class, such as Least-Squares Regressions, Regularized Least-Squares, and Truncated Singular-Value Decomposition, to determine which factors contribute most to the perceived happiness of a country. To do so, we downloaded 2 different "happiness indices" (numbers that prior researchers associated with different countries during different years as a measure of how happy the countries are) and we trained multiple different regression models to predict these indices and a composite index using 18 potential "quality-of-life indicators" for 100 countries from 2014-2016. The weights of our regression model with the highest magnitudes correspond to the most significant features, and, therefore, the indicators that contribute the most to a country's perceived quality of life.

Everyone contributed to all aspects of this project, though my main contributions to this project are writing the code for performing each of the different regressions and writing the Methods section of the paper.

Our final report is embedded below. Please contact jserf02@gmail.com for the code that produced these regressions or for the final processed data points.

Paper