Machine learning for predicting economic growth: How can Zimbabwe achieve Vision 2030?

Abstract

Though Machine Learning has been around for a while, it is still considered a new tool for economists and in its application to predicting economic growth. Studies that apply machine learning to predicting economic growth have found that the Random Forest algorithm is currently the best performing machine learning algorithm for predicting economic recessions and economic growth. However, besides studies evaluating the various machine learning algorithms, there is limited literature on the application of these techniques to help economists and policymakers solve problems. Developing African countries, like Zimbabwe, with their unique economic growth challenges, can harness the predictive qualities of this technology in development planning, setting, and achieving growth targets. In this thesis, I apply the random forest algorithm to make income predictions for Zimbabwe, in the country's hopes to attain upper-middle-income status by 2030.