UnFEAR: Unsupervised Feature Extraction Clustering with an Application to Crisis Regimes Classification

24 Pages Posted: 18 Feb 2021

See all articles by Jorge A. Chan-Lau

Jorge A. Chan-Lau

International Monetary Fund (IMF) - International Capital Markets Department

Ran Wang

University of California Riverside

Multiple version iconThere are 2 versions of this paper

Date Written: November 25, 2020

Abstract

We introduce unFEAR, Unsupervised Feature Extraction Clustering, to identify economic crisis regimes. Given labeled crisis and non-crisis episodes and the corresponding features values, unFEAR uses unsupervised representation learning and a novel mode contrastive autoencoder to group episodes into time-invariant non-overlapping clusters, each of which could be identified with a different regime. The likelihood that a country may experience an econmic crisis could be set equal to its cluster crisis frequency. Moreover, unFEAR could serve as a first step towards developing cluster-specific crisis prediction models tailored to each crisis regime.

Keywords: clustering, unsupervised feature extraction, autoencoder, machine learning, deep learning, biased label problem, crisis prediction

JEL Classification: C1, C45, G1

Suggested Citation

Chan-Lau, Jorge Antonio and Wang, Ran, UnFEAR: Unsupervised Feature Extraction Clustering with an Application to Crisis Regimes Classification (November 25, 2020). Available at SSRN: https://ssrn.com/abstract=3773093 or http://dx.doi.org/10.2139/ssrn.3773093

Jorge Antonio Chan-Lau (Contact Author)

International Monetary Fund (IMF) - International Capital Markets Department ( email )

700 19th Street NW
Washington, DC 20431
United States

Ran Wang

University of California Riverside ( email )

900 University Avenue
Riverside, CA 92521
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
3
Abstract Views
56
PlumX Metrics