Hidden Markov model (HMM)

A statistical model that guesses which hidden state, or regime, a system is in by looking at data it can see, such as returns.

Why it matters

Market regimes cannot be observed directly. An HMM estimates them from patterns in returns and volatility, along with the chance of switching from one to another.

How Ocarina Research shows it

Signals uses an HMM fit on daily returns to label market regimes.

Example

Sample numbers for illustration, not real data

Given 5 years of daily returns, an HMM might estimate that a calm regime lasts about 60 trading days on average before switching.

Sources

  • James D. Hamilton, A New Approach to the Economic Analysis of Nonstationary Time Series and the Business Cycle, Econometrica 57(2), 1989

All terms

We use cookies and similar technologies to make Ocarina Research work and to understand how it's used. You can accept all, reject non-essential, or customize. See our Privacy Policy for details.