An Analysis of Machine Learning Techniques for Economic Recession Prediction

Author

Sheridan Kamal, The Graduate Center, City University of New YorkFollow

Date of Degree

6-2021

Document Type

Capstone Project

Degree Name

M.S.

Program

Data Analysis & Visualization

Advisor

Howard Everson

Keywords

Economic Recession, Machine Learning, Support Vector Classifier

Abstract

In this project I used the supervised machine learning methods logistic regression, decision tree classifier, k nearest neighbor classifier, and support vector classifier, to determine the best method to predict economic recessions. To do this, I used the function train_test_split to create training and testing sets and the function TimeSeriesSplit to create walk-forward cross validation sets to use when tuning the model parameters. Each machine learning method was trained on both scaled and unscaled data and was performed using default parameters and using the tuned parameters so that there were four models of each method. It was determined that the tuned Support Vector Classifier model trained on scaled data with a precision-recall area under the curve (PR AUC) score of 0.83 was the optimal model to predict economic recessions.

Recommended Citation

Kamal, Sheridan, "An Analysis of Machine Learning Techniques for Economic Recession Prediction" (2021). CUNY Academic Works.
https://academicworks.cuny.edu/gc_etds/4378