Welcome to the UPF Digital Repository

The wisdom of the crowd: using ensemble machine learning techniques as an early warning indicator for systemic banking crises

Show simple item record

dc.contributor.author Lavagna, Gabriela
dc.contributor.author Patterson, Helena
dc.contributor.author Razmadze, Robizon
dc.date.accessioned 2021-12-09T10:42:44Z
dc.date.available 2021-12-09T10:42:44Z
dc.date.issued 2021-07
dc.identifier.uri http://hdl.handle.net/10230/49160
dc.description Treball fi de màster de: Master's Degree in Economics and Finance. Finance programme. Curs 2020-2021
dc.description Director: Dmitry Kuvshinov
dc.description.abstract We develop early warning models for systemic crises prediction using machine learning techniques on macrofinancial data for 36 countries for quarterly data spanning 1970-2013. Machine learning models outperform logistic regression in out-of-sample predictions under the recursive window forecasting mechanism. In particular, using the ensemble random forest algorithm for both feature selection and prediction substantially outperforms the logit models. We identify the key economic and financial drivers of our models using the random forest framework by extracting each feature’s Gini impurity and corresponding information gain. Throughout the time period, the most important predictors are credit, foreign liabilities, asset prices and foreign currency reserves.
dc.description.abstract Desarrollamos modelos de alerta temprana para la predicción de crisis sistémicas utilizando técnicas de machine learning utilizando variables macro financieras de 36 países con datos trimestrales que abarcan el período 1970-2013. Los modelos de machine learning superan a la regresión logística en predicciones fuera de la muestra utilizando ventanas recursivas. En particular, el uso de random forest, tanto para la selección de variables como para predicción, supera sustancialmente a los modelos logit. Identificamos a los impulsores económicos y financieros clave de nuestros modelos extrayendo la impureza de Gini de cada variable y la correspondiente ganancia de información. A lo largo de nuestra muestra, los predictores más importantes son el crédito, los pasivos externos, los precios de activos y las reservas internacionales.
dc.format.mimetype application/pdf
dc.language.iso eng
dc.rights Atribución-NoComercial-SinDerivadas 3.0 España
dc.rights.uri http://creativecommons.org/licenses/by-nc-nd/3.0/es/
dc.subject.other Treball de fi de màster – Curs 2020-2021
dc.title The wisdom of the crowd: using ensemble machine learning techniques as an early warning indicator for systemic banking crises
dc.type info:eu-repo/semantics/masterThesis
dc.subject.keyword Financial crises
dc.subject.keyword Machine learning
dc.subject.keyword Early warning models
dc.subject.keyword Crisis financieras
dc.subject.keyword Modelos de alerta temprana
dc.rights.accessRights info:eu-repo/semantics/openAccess

Thumbnail

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account

Statistics

In collaboration with Compliant to Partaking