Welcome to QSignature’s documentation!

QSignature is a model-free dynamical regime classification framework for causal time series data.

The framework provides two persistence timescale estimators — the signed centroid τ_s and the unsigned centroid τ_u — that classify dynamical regimes without parametric assumptions.

For theoretical background, see:

• QSignature 1.0 Framework : IEEE ISDFS 2026 :cite:`qsignature2026_isdfs`

• Theorems for Environmental Signature : Research Square :cite:`qsignature2026_theorems`

Getting Started

• Installation
  • Using pip
  • From source
  • Dependencies
  • Verification
• Quick Start
  • Script 1: Basic Classification (10 lines)
  • Script 2: Growth vs Decay Detection
  • Script 3: QSpace Classification Demo
  • Understanding R_su
  • Important Notes
  • Next Steps

Core Concepts

• Synthetic Data Generation
  • Basic Usage
  • Available Systems
  • Adding Noise
  • Sampling Irregularities
  • Batch Dataset Generation

Reference

• API Reference
  • Core Estimators
  • Diagnostic Ratios
  • High-Level Functions
  • Lambda and Classification
  • Classes
• Examples
  • Example 1: Quick Start
  • Example 2: Growth vs Decay Detection
  • Example 3: QSpace Classification
  • Example 4: Noise Reduction with QSmooth
  • Example 5: Synthetic Data Generation
• Frequently Asked Questions
  • What is QSignature?
  • Does QSignature require model fitting?
  • What types of data can QSignature analyze?
  • How do I interpret R_su?
  • What is Λ (envelope growth rate)?
  • Why is τ_u phase-invariant?
  • How do I cite QSignature?
  • Where can I find examples?
  • What if my data is noisy?

Community

• Acknowledgments
  • DeepSeek
  • Contributors
  • Citation

Indices and tables

• Index

• Module Index

• Search Page