Loan Portfolio Engines - Challenges and Risk Classification

Explore a comprehensive conference talk on loan portfolio engines presented by Stanford's Kay Giesecke at Conf.Startup.ML. Delve into typical questions and challenges in the field, including those specific to Asset-Backed Securities (ABS). Learn about risk classifiers and their underlying mechanisms, as well as time-varying factors and correlation in loan portfolios. Examine the transition function and its applications through practical examples. Evaluate predictive performance and understand pool-level modeling techniques, including limit laws and second-order approximations. Discover computational considerations and strategies for optimizing loan pools, with real-world examples to illustrate key concepts. This in-depth presentation offers valuable insights for professionals and researchers in the fields of machine learning, finance, and risk management.

Introduction
Typical Questions
Challenges
Challenges of ABS
Risk Classifier
Under the Hood
Time Varying
Feedback Factors
Correlation
Transition Function
Examples
Predictive Performance
Pool Level Modeling
Limit Law
Second Order Approximation
Computational Times
Optimization of Pools
Example