Hidden Markov Models for Inferring Stock Market Regimes
Summary
This introductory explanation builds from the Markov chain assumption: the next state depends on the current state rather than the full history. It then describes a hidden Markov model (HMM), where observable data are probabilistically linked to an unobserved state sequence. A dice example illustrates hidden states, transition probabilities between them, and emission probabilities that connect each hidden state to an observed outcome.
Applied to stocks, the observable sequence might include prices and trading volume, while the hidden states represent conditions such as bull, bear, range-bound, or rebound markets. The note distinguishes three common tasks: infer the hidden state sequence when model parameters are known, calculate the likelihood of observations, and estimate unknown model parameters from observations. It says market applications require state decoding and parameter estimation, but does not develop those algorithms or provide a worked stock example, validation, or evidence of forecasting performance.
Key ideas
- A Markov chain models transitions whose next state depends on the current state.
- An HMM links observable data to hidden states through emission probabilities and state transitions.
- In a market example, prices and volume are observable while the market regime is latent.
- HMM tasks include decoding hidden states, evaluating observation likelihood, and estimating unknown parameters.
- The note is conceptual and does not show a complete market implementation or performance evaluation.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.