Member-only story
During the past week or so I have been studying quantile regression, which is a variation of linear regression. The most recent piece that I have written about quantile regression can be found here:- https://medium.com/@tracyrenee61/does-sklearns-quantile-regression-perform-better-than-linear-regression-a8eaaff52c4d
I have also made a code review using tensorflow to create a quantile regression deep learning model, and it can be viewed here:- https://youtu.be/d8cC2wRo1mY
I decided therefore to read some research papers on the subject, and one that I have found is the paper entitled, “Deep Quantile Regression” by King’s College London, and it can be found here:- https://www.kcl.ac.uk/business/assets/pdf/dafm-working-papers/2021-papers/deep-quantile-regression.pdf
The authors of this work proposed a deep quantile estimator, using neural networks to examine a non-linear association between the conditional quantiles of a dependent variable and predictions. They used a deep quantile estimate to forecast Value at Risk (VaR) and find significant gain over linear regression alternatives.
Value at Risk (VaR)
Value at Risk (VaR) is a risk management metric used to estimate the maximum potential loss of an investment or portfolio over a specific time horizon and at a certain level of confidence. VaR is commonly used in finance and investment to assess and quantify the risk associated with different assets or portfolios.
