Consumer Bankruptcy Prediction Using Balanced and Imbalanced Data

This paper examines the usefulness of logit regression in forecasting consumer bankruptcy households using an imbalanced dataset. The research on prediction is paramount importance as it aims to build statistical models that can identify consumers a difficult financial situation may lead bankruptcy. In face current global pandemic crisis, future household finances uncertain. change macroeconomic and microeconomic requires searching for better more precise methods. relies four samples households: two learning (imbalanced balanced) testing from Survey Consumer Finances (SCF) which was conducted United States. results show predictive performance model based balanced sample effective compared one sample. Furthermore, mortgage debt assets ratio, age, being married, having credit constraints, payday loans or payments than 60 days past due last year appear be predictors increase risk becoming bankrupt. Moreover, both ratio card overall owning house decrease going