Machine Learning for Financial Risk Management with Python

Afterword

This book tries to show how ML and deep learning models can be incorporated to address financial problems. This, of course, does not mean that the book includes all models with all necessary steps required to deploy models in the industry, but I have tried to focus on the topics I felt would be of the greatest interest to the reader.

Recent developments in AI indicate that almost all traditional financial models are outperformed by AI models in terms of predictive performance, and I believe that adopting these models and having improved predictive performance in the industry will help finance practitioners make better decisions.

However, despite these developments and recent hype around AI, the deployment level of AI models is not even close to where it should be. The opaque nature of these models stands out as the first and foremost reason. However, there are constant and tremendous improvements toward getting more explainable AI models, and these orchestrated efforts have started to pay off, as Prado has argued, whether ML is opaque or not depends on the person using it, not on the ML algorithms themselves.

Thus, in my opinion, the long tradition of using parametric models as well as resistance to paradigm shift are the primary reasons for the slow and reluctant adoption of AI models.

Hopefully, this book paves the way for embracing AI models and provides a smooth and convenient transition to using them.