Empirical Mode Decomposition (EMD)

Parameters: num_imfs = 5 (3–10) • sifting_threshold = 0.001 (0.0001–0.01)

Overview

Empirical Mode Decomposition (EMD) is an advanced signal processing technique developed by Norden E. Huang in 1998 as part of the Hilbert-Huang Transform (HHT). Unlike traditional methods like Fourier analysis that assume data is linear and stationary, EMD adaptively decomposes complex price movements into a series of intrinsic mode functions (IMFs), each representing a different oscillatory mode within the data. This makes EMD particularly powerful for financial markets, where price behavior is inherently non-linear, non-stationary, and contains multiple overlapping cycles that traditional indicators cannot detect.

The method uses an iterative sifting process to extract IMFs directly from the data without imposing any predetermined basis functions. Each IMF represents a simple oscillatory mode with a characteristic timescale - from high-frequency noise to long-term trends. The decomposition follows: Original Signal = IMF₁ + IMF₂ + ... + IMFₙ + Residue, where IMF₁ contains the highest frequency components (market noise), middle IMFs capture trading cycles, and the residue represents the underlying trend. This multi-scale analysis reveals hidden periodicities and market structures that are crucial for understanding complex market dynamics.

Interpretation & Trading Signals

IMF Component Analysis:

  • IMF₁ (Highest Frequency): Market noise, scalping opportunities, intraday volatility
  • IMF₂-₃ (High Frequency): Short-term trading cycles, swing patterns
  • IMF₄-₅ (Mid Frequency): Medium-term market cycles, position trading
  • Residue (Trend): Primary trend direction, long-term investment signals

Trading Applications:

  • Cycle Detection: IMF zero crossings indicate cycle turning points
  • Multi-Scale Trading: Use different IMFs for different trading timeframes
  • Phase Alignment: When multiple IMFs align, expect major moves
  • Portfolio Optimization: Analyze cross-correlations among IMFs for diversification

Key Considerations:

  • Non-Causal Nature: EMD requires complete data, challenging for real-time trading
  • End Effects: Recent data points may be less reliable due to boundary conditions
  • Mode Mixing: Sometimes different scales mix in one IMF, requiring careful interpretation
  • Market Similarity: Studies show EMD reveals similar patterns across different global markets

Example Usage

Code examples will be available once the Rust implementation is complete.

Performance Analysis

Related Indicators

Bandpass Filter

Technical analysis indicator

Correlation Cycle

Technical analysis indicator

Detrended Oscillator

Technical analysis indicator

Decycler

Technical analysis indicator

Hilbert Transform - Dominant Cycle Period

Technical analysis indicator

Hilbert Transform - Dominant Cycle Phase

Technical analysis indicator