Small and medium-sized enterprises’ time to default: an analysis using an improved mixture cure model with time-varying covariates
The authors put forward a method using a support vector machine to enhance the exploration of nonlinear covariate effects if SMEs never default while also considering time-varying and fixed covariates for the incidence and latency of an event.
The authors explore possible instabilities in applying Cox PH models and conduct numerical studies to demonstrate the same linear specification error from APC models an occur in Cox PH estimation.
This paper uses empirical methods to investigate how psychometric data can be used to augment traditional credit models.
Based on a comprehensive sample, the authors benchmark machine learning models in the prediction of financial distress of publicly traded US firms, with gradient-boosted tress outperforming other models in one-year-ahead forecasts.