I am releasing a draft paper on stress, reverse stress, and risk measure testing. The paper is produced by a jupyter notebook (thus reproducible and easily adaptable to other cases). We are thinking about open-sourcing our Machine Learning framework.
There are much more to write on this topic that what this paper present as data quality, time-dependent PnL functions, etc etc. This could lead us to a mid-term research project. However I was wondering if it is worthwhile to spend more time on this topic. If yes, what are the priority topics to look to ?
Any thoughts welcome !