The application of Machine Learning methods in Economics and Econometrics

The contribution of the thesis is to compute a survey of literature showing the main Machine Learning algorithm, how they worked and how they can be classified. The main classification describes are Supervised and Unsupervised Machine Learning, parametrized and non-parametrized tests. Then different regularizer depending on the function class. Subsequently, specification of different regression, in particular Lasso Regression, Ridge Regression, Elastic Net, Regression trees and finally Neural Network. The work continues by pointing to the best known techniques for improving the performance of the algorithm, SGD, Boosting, Bootstrap, Bagging, Bumping, Orthogonalization, Cross- validation. Following the table of contents, there are the specification of advantages and disadvantages coming from both Machine Learning and traditional methods, as the OLS method. Lastly, an examination of different real cases in which the algorithms of Machine Learning are applied. The main area that are selected: Poverty, Banking and Finance and Politics and Policy. The thesis provides an overview of Machine Learning and how it can be applied in economics and econometrics, drawing tangible cases.