Examples of Machine Learning Uses in Quantitative Finance

Summary

The document responds to a request for credible examples of machine learning in quantitative finance. It emphasizes that machine learning methods are general-purpose tools that can be applied to financial, economic, or other data, while noting that many strong financial applications are proprietary and therefore difficult to inspect. The examples mentioned include deep learning for mortgage risk estimation, finance-focused public data challenges, an automated market-bubble predictor, large-scale machine learning tools, and a company that applied data analysis to hedge funds.

These references illustrate possible uses in risk estimation, prediction, and research infrastructure, but the document does not describe model designs, datasets, evaluation procedures, or measured trading performance. It therefore serves as a pointer to areas and sources for further study rather than evidence that machine learning produces reliable trading profits. The brief mention of another research repository likewise supplies no assessment of the cited work. Any conclusions about predictive value, robustness, or live deployment would require examining the underlying studies and their validation methods.

Key ideas

• Machine learning methods can be applied to many kinds of data, including financial data.
• Proprietary finance applications can make credible examples difficult to inspect.
• The examples span mortgage risk estimation, bubble prediction, data challenges, and machine learning infrastructure.
• The document provides pointers but no model specifications or performance evidence.
• A trading claim requires reviewing the underlying study and its validation approach.

Tags

# Can someone please share examples of machine learning in quantitative finance?


# Can someone please share examples of machine learning in quantitative finance?












There has been a lot said about the application of AI, ML and Neural Networks in trading for predictive modelling. I was unable to find any relevant examples that prove a credible output based on these applications.

## Answer by DJohnson (score 2)

https://quant.stackexchange.com/a/40740

AI, ML and NNs are generic algorithms or methods which leverage data agnostically. In other words, they can be deployed with data of any type be it digital, economic, financial, whatever. That said, most of the good examples are proprietary, hence your difficulty in identifying specific models.

1) Estimating risk is an essential element in finance. This paper Deep Learning for Mortgage Risk is a good example of that: https://arxiv.org/pdf/1607.02470.pdf

2) Kaggle has had some financial challenges: https://www.kaggle.com/tags/finance

3) Didier Sornette's Financial Crisis Observatory has links to his market bubble predictor model which is, to a large extent, automated. http://www.er.ethz.ch/financial-crisis-observatory.html

4) Google has many tools for ML including Sibyl: Google’s system for Large Scale Machine Learning (https://www.kdnuggets.com/2014/08/sibyl-google-system-large-scale-machine-learning.html) and Tensorflow (https://www.tensorflow.org/).

5) Ufora.com was a good example, but it appears to be out of business. It's business model could be a template for future efforts of a similar nature. http://time.com/money/3890808/hedge-funds-data-analysis-ufora/

## Answer by Giladbi (score 0)

https://quant.stackexchange.com/a/40739

did you try to search SSRN or anything else like this? https://arxiv.org/abs/1807.02787

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.