---
title: Spatio-Temporal Machine Learning for Mortgage Credit Risk
url: https://www.ml-quant.com/papers/ssrn/4975440/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: SSRN 4975440
source_url: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4975440
featured: 2024-10-09
citations: unknown
topic: Risk, Credit & Banking
---


# Spatio-Temporal Machine Learning for Mortgage Credit Risk

A new machine learning model for credit risk combines tree-boosting with a latent spatiotemporal Gaussian process model, offering more accurate predictions of default probabilities and loan portfolio losses.

- Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4975440
- Identifier: SSRN 4975440
- Released: 2024-10-03
- First featured: Quant Letter No. 69 (2024-10-09): https://www.ml-quant.com/issues/2024-10-09/
- Citations (Semantic Scholar): not tracked
- Published in: not yet
- Topic: Risk, Credit & Banking

## Related

- [Climate Risks in Real Estate](https://www.ml-quant.com/papers/ssrn/5120353/): The study shows how physical climate risks, specifically river floodings, can affect the credit risk parameters and internal capital calibration of banks.
- [Upper Comonotonicity and Risk Aggregation Under Dependence Uncertainty](https://www.ml-quant.com/papers/arxiv/2406.19242/): The research investigates the concept of dependence uncertainty and its effect on tail risk measures in relation to credit risk, showing that even minor positive dependence between losses can lead to perfectly correlated tails beyond a certain point.
- [The Relative Entropy of Expectation and Price](https://www.ml-quant.com/papers/arxiv/2502.08613/): The article explores the non-linear pricing in incomplete securities markets, measuring strategic risks using an entropic risk metric and adjusting the price for market incompleteness and default risk.
- [Modelling the term-structure of default risk under IFRS 9 within a multistate regression framework](https://www.ml-quant.com/papers/arxiv/2502.14479/): A study comparing three loan behavior modeling techniques finds multinomial logistic regression to be the most effective, potentially improving loss reserve estimates in banking.
- [A Markov approach to credit rating migration conditional on economic states](https://www.ml-quant.com/papers/arxiv/2403.14868/): A novel credit rating migration model has been developed, incorporating economic state changes and utilizing Markov theory for various rating philosophies analysis.
- [Recalibrating binary probabilistic classifiers](https://www.ml-quant.com/papers/arxiv/2505.19068/): The article discusses recalibrating binary probabilistic classifiers from a distribution shift perspective, introducing two new methods for conservative results in credit risk assessments.