Approx. 5,000 mortgage loans and 1,500 defaults
with workout experiences
- The mortgage data includes 15,000 defaults with workout losses for 50,000 mortgages observed and 60 quarters. The data also provides many risk controls, payoff events and exposures for feature engineering.
- The code includes the dcr.py (for Python) and dcr.R module (for R) which is a collection of functions used in the book to accelerate your learning.
- All codes included in the book may be copied and pasted for the Amazon Kindle Version. You would need to install the Kindle App on your machine and delete spaces that occasionally are inserted before variables names in quotation marks.
- Please cite "Roesch, D., & Scheule, H. (2021). Deep Credit Risk: Machine Learning with Python, Kindle Direct Publishing" or "Scheule, H. & Roesch, D. (2022). Deep Credit Risk: Machine Learning in R, Kindle Direct Publishing" when using the data and code.