by Joseph Mikael, Simon Fécamp, Xavier Warin
Continue on SelfMadeQuant
Suggested foundations for financial machine learning and risk measurement and stress. Specialist prerequisites vary by paper.
Preserve time order and avoid future information in training.
Establish an interpretable baseline before a flexible model.
Represent losses as a distribution rather than one forecast.
Distinguish a loss quantile from the average loss in the tail.
Account for dependence and overlapping observations.
Examine which information contributes to predictions.
Compare risk estimators and their assumptions.
Choose a reproducible strategy, specify the trading rule, and test what remains after costs and estimation uncertainty.
Connect diversification, common risk factors, investor views, and the properties of a risk measure.