Reference Guide#
On these pages you can find documentation for infomeasure.
The package implements a comprehensive suite of information-theoretic measures—such as Entropy (H),
Mutual Information (MI), Conditional MI, Transfer Entropy (TE),
Conditional TE, and Cross-Entropy—for
both discrete and continuous-valued data.
The package provides multiple estimation techniques for each measure, including bias-corrected estimators, Bayesian approaches, coverage-based methods, and specialized techniques for different data characteristics. For discrete data, estimators range from a simple maximum likelihood to sophisticated methods like NSB, Miller-Madow, and Chao-Wang-Jost. For continuous data, methods include kernel density estimation and the Kraskov-Stögbauer-Grassberger algorithm.
For guidance on selecting the appropriate estimator for your data, see the Estimator Selection Guide. For detailed information on programmatic usage and API details, please refer to the API Reference.
- Introduction
- Estimation
- Estimator Usage
- Estimator Selection Guide
- How to Use This Guide
- Start Here: What Type of Data Do You Have?
- Discrete Data Selection
- Continuous Data Selection
- Time Series Data Selection
- Data Type Help
- Information Measure Selection Guide
- Performance Considerations
- Practical Examples
- Quick Decision Summary
- Time Lag Selection for Transfer Entropy and Mutual Information
- Additional Information
- Entropy (H)
- Conditional H
- Cross-Entropy
- Mutual Information (MI)
- Conditional MI
- Transfer Entropy (TE)
- Conditional TE
- Statistical Tests
- Kullback–Leibler Divergence (KLD)
- Jensen–Shannon Divergence (JSD)
- Settings