Λογότυπο αποθετηρίου

Πλοήγηση ανά Συγγραφέα "Giannakopoulos, Dimitrios"

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
Φιλτράρετε τα αποτελέσματα πληκτρολογώντας τα πρώτα γράμματα
Τώρα δείχνει 1 - 1 από 1
  • Μικρογραφία εικόνας
    Τεκμήριο
    Integration of machine learning and econometric approaches for loan performance analysis: a comparative study
    (2024-03-29) Giannakopoulos, Dimitrios; Γιαννακόπουλος, Δημήτριος; Athens University of Economics and Business, Department of Economics; Tzavalis, Elias; Pagratis, Spyridon; Dendramis, Yiannis
    Assessing credit risk through machine learning typically involves the application of classification algorithms to distinguish between reliable and unreliable customers based on historical data. This thesis delves into the application through extensive literature review of machine learning techniques for credit risk assessment within the banking sector, highlighting the shift from traditional statistical methods to advanced AI-driven algorithms due to their efficiency in handling complex datasets. In extension, the research applies various machine learning models, including logistic regression, decision trees, and random forests, to an unbalanced dataset to assess their impact on predicting loan defaults, AND also explores the use of SMOTE for dataset balancing, aiming to improve model performance in predicting financial outcomes In the analysis of imbalanced datasets, tree-based methodologies demonstrate a marginal superiority over logistic regression as ordinal classifiers. However, logistic regression distinguishes itself with superior discriminative power, as evidenced by higher Area Under the Curve (AUC) score values, across both balanced and imbalanced datasets.