Μεταπτυχιακές Εργασίες

Μόνιμο URI για αυτήν τη συλλογήhttps://pyxida.aueb.gr/handle/123456789/7

Περιήγηση

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 1 από 1

Behavioral scorecard using machine learning techniques
(2021) Anastasopoulos, Nikolaos; Αναστασόπουλος, Νικόλαος; Athens University of Economics and Business, Department of Informatics; Karlis, Dimitrios; Louridas, Panagiotis; Vassalos, Vasilios
In the context of credit scoring, behavioral scorecards are used by financial institutions over time, in order to monitor the performance of their existing clients. Scorecards are used to generate clients’ scores based on their behavior in relationship with their financial institution. For such a crucial decision, past demographic and financial data of clients (behavioral characteristics) are important to be collected so as to build an automated behavioral score prediction model, based on a machine learning classifier or statistical models using machine learning techniques. The present work, focuses on collecting and sampling the appropriate data, cleaning, and performing the necessary preprocessing steps, feature selection using univariate and supervised wrapper techniques, and model development, based on baseline classifiers (Logistic Regression, Decision Trees, Random Forest). Traditional approaches of Weight of Evidence and Information Value criterions for credit scoring are also examined. After analysis, the best performing classifier will be used in order to create a scorecard that will be able to generate scores based on clients’ characteristics.