Learning Multinomial Logit model (enhanced MNL with ANN) Master thesis research project "Interpretable neural networks for a multiclass assessment of credit risk". Based on Sifringer et al. (2020) L-MNL model.
Contents:
-
Word document.
-
Python files (uses Tensorflow and Keras libraries):
-
run_lending_main: main file, imports and pre-processes the data, runs a 5-Fold Stratified Cross-Validation, evaluates performance measures (AUC, accuracy, F1 score, MAE, and BIC), and extracts corresponding coefficients and standard deviations for the MNL and L-MNL model.
-
models: contains the L-MNL model definition
-
grad_hess: functions for extracting betas (coefficients) and corresponding standard deviations of the L-MNL model.
-
-
R files:
- credit_lending: runs the Cumulative logit model, and saves the predictions in a csv file. Evaluates the Friedman test and does a pairwise comparison using Wilcoxon rank sum test for the performance results of all models.
References:
Sifringer, B., Lurkin, V., & Alahi, A. (2020). Enhancing discrete choice models with representation learning. Transportation research. Part B: methodological, 140, 236-261. https://doi.org/10.1016/j.trb.2020.08.006
George, N. (2019, April 10). All Lending Club loan data, Version 3. Retrieved February 2020 from https://www.kaggle.com/wordsforthewise/lending-club.