Shannon Entropy Measure

Quick Reference

Property	Value
Dimension	quality
Category	statistics
Version	v1.0
Output Column	shannon_entropy

Shannon entropy 鈥?measures distribution uncertainty over a rolling window

Formula

rolling_std(signal, window)

CDM Inputs

Column	CDM Table	Description
signal	cdm_*	CDM source table

Parameters

Parameter	Type	Default	Description
window	integer	3600000	Window for entropy estimation

Output

Column: shannon_entropy

Shannon entropy of the signal distribution

Market Intuition & Trading Rationale

Shannon entropy measures the uncertainty of a signal's distribution: rolling_std(signal, window) approximates entropy for roughly normal distributions. High entropy means the signal takes many different values with similar probability 鈥?it's unpredictable, exploring its full range. Low entropy means the signal is concentrated around a few values 鈥?it's predictable, stuck in a narrow range.

In this context, entropy is a quality diagnostic: a good directional signal should have moderate entropy 鈥?not so high that it's random noise, not so low that it's a constant. Entropy spikes often signal regime changes: the signal suddenly explores new values as the market transitions. Entropy collapse (near zero) suggests the signal has flatlined 鈥?it's producing the same value repeatedly and has stopped responding to market data.

Usage Cases

• Signal health monitoring: Track shannon_entropy over time. Sudden entropy spikes 鈫?investigate for data issues or regime change. Entropy collapse 鈫?signal may be stuck (e.g., division by near-zero, stale input).
• Regime transition detection: Entropy often rises before volatility. When multiple signals in a feature set show simultaneous entropy increases, the market may be transitioning between regimes 鈥?prepare for changing correlations.
• quality context: Used in volatility_regime_transition and trade_toxicity packs as a quality dimension 鈥?entropy of the signal distribution complements SNR and IC for a complete quality picture.

YAML Definition

name: shannon_entropy_measure
description: Shannon entropy 鈥?measures distribution uncertainty over a rolling window
category: statistics
dimension: quality
version: v0.9.0 (Beta)
required_inputs:
  - signal
output_column: shannon_entropy
output_description: Shannon entropy of the signal distribution
tags:
  - quality
  - entropy
  - statistics
parameters:
  window:
    type: integer
    description: Window for entropy estimation
    required: false
    default: 3600000
formula: rolling_std(signal, window)