PREDICTING EXTERNAL MACROECONOMIC CRISES: MACHINE LEARNING PERSPECTIVE

Date

2024-3-06

Author

CELIK, Songul Siva

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

91
views

76
downloads

Cite This

In this study, our primary objective is to predict Sudden Stops using Machine Learning (ML) methods and evaluate their out-of-sample prediction power. Conducted in two phases, the first involves establishing a baseline with Forbes and Warnock (2021) as our model, replicating and assessing its out-of-sample predictions. We then introduce various ML methods, including Elastic Net, Random Forests, Support Vector Machines, kNN, AdaBoost, XGBoost, and Multi-Layer Perceptron, for a comprehensive comparison against the baseline, utilizing metrics like accuracy, precision, recall, F1-score, and ROC curve. In the second phase, we expand the dataset from the IMF, prioritizing data availability, and employ ML methods for feature selection. Selected features are used for ML estimation, involving traditional methods like Elastic Net, Random Forests, Support Vector Machines, XGBoost, Logistic Regression, and the modern deep learning method, a Long Short-Term Memory (LSTM).The analysis aims to explore potential improvements in out-of-sample performances facilitated by ML algorithms and feature selection.

Subject Keywords

Machine Learning, Sudden Stops, Deep Learning, Out of Sample Prediction

URI

https://hdl.handle.net/11511/109045

Collections

Graduate School of Social Sciences, Thesis