Supervisor

Taufique Ahmed

Programme

HDIP in Data Analytics for Business

Subject

Computer Science

Abstract

Economic recessions can have significant consequences for businesses, households, financial markets, and policymakers, creating a need for improved understanding of the factors associated with economic downturns. This project analyses historical US economic and financial data to examine patterns associated with previous recessionary periods and assess whether machine learning can be used to identify recession risk. The study uses a two-tier analytical approach, combining a macroeconomic dataset covering 1960 to the present with a more detailed dataset from 1987 onwards that incorporates financial markets, house prices, and consumer confidence. Building on previous exploratory analysis, the project develops and evaluates three machine learning models, Logistic Regression, Random Forest, and XGBoost, using engineered features derived from historical economic indicators. The analysis examines the relationships between variables such as unemployment, interest rates, industrial production, inflation, financial markets, and consumer confidence and their association with recessionary periods. Model performance is evaluated using appropriate classification metrics and time-based validation. The findings provide insights into historical recession patterns and demonstrate how machine learning can support data-driven analysis of recession risk, while recognising the limitations associated with changing economic conditions and the relatively small number of historical recession periods.

Date of Award

2026

Full Publication Date

2026

Access Rights

open access

Document Type

Capstone Project

Resource Type

thesis

Download

Share

COinS