How’s the market today?
This fair question has an implicit assumption:
Markets
DON’T
behave
exactly
the same way
every single day.
But they’re
HIDDEN
HMM involves two interconnected models
All models are wrong, but some are useful. (George E. P. Box)
HMM may be interpreted as an extension of a mixture model
Common sense check: Observing small returns makes us more confident about using the low volatility model.
Correlation in belief states is higher for stocks in the same industry (compare F/GM vs F/AIR).
What if hidden states were hierarchical?
Risk state of a global portfolio may be mapped into components
\(\text{Country} + \text{Industry} + \text{Individual}\)
I’m currently exploring
Hierarchical Hidden Markov Models applied to finance
for my Google Summer of Code 2017 project.
Check the extended material for this talk
(R Notebook, fully working code, references and neat plots)
https://github.com/luisdamiano/rfinance17