Dow Jones Hidden Markov Model Analysis

A comprehensive project for analyzing Dow Jones financial time series data using Hidden Markov Models (HMM) and the Viterbi algorithm.

Overview

This project:

1. Fetches historical Dow Jones (^DJI) data from 1995 to present
2. Preprocesses and engineers financial features (returns, volatility, momentum)
3. Builds a Hidden Markov Model with multiple hidden states
4. Applies the Viterbi algorithm to decode the most likely hidden state sequence
5. Visualizes results and analyzes market regimes

Project Structure

├── src/
│   ├── data_fetcher.py          # Download Dow Jones data
│   ├── feature_engineering.py   # Create financial features
│   ├── hmm_model.py             # Hidden Markov Model implementation
│   ├── viterbi_algorithm.py     # Viterbi algorithm for state decoding
│   └── visualization.py         # Plotting and analysis
├── data/                        # Downloaded data and results
├── main.py                      # Main execution script
├── requirements.txt             # Project dependencies
└── README.md                    # This file

Installation

1. Clone or download this project
2. Create a virtual environment (optional but recommended):

   python -m venv venv
   venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

Usage

Run the main analysis:

python main.py

This will:

• Download Dow Jones historical data
• Train the HMM model
• Run Viterbi algorithm
• Generate visualizations
• Save results to the data/ folder

Key Components

Data Fetcher

Downloads historical Dow Jones data using yfinance library.

Feature Engineering

Calculates returns, volatility, momentum, and other technical indicators.

Hidden Markov Model

Implements HMM with:

• Multiple hidden states (representing market regimes: Bull, Normal, Bear)
• Gaussian emission probabilities
• Estimated transition and emission matrices

Viterbi Algorithm

Decodes the most likely sequence of hidden states given the observed returns data.

Results

The model identifies different market regimes and their transitions, helping to understand:

• Bull market states (high returns, low volatility)
• Normal market states (moderate returns)
• Bear market states (negative returns, high volatility)

Requirements

• Python 3.7+
• See requirements.txt for dependencies