Ulteriori informazioni
Beginning with a gentle introduction to causal discovery and the foundations of probability and statistics, this textbook is written in a highly pedagogical way. By uniting probability theory, statistical inference, and graph theory, the book offers a systematic pathway from foundational principles to cutting-edge algorithms, including independence tests, the PC algorithm, LiNGAM, information criteria, and Bayesian methods. Far more than a theoretical treatment, this volume emphasizes hands-on learning through Python implementations, carefully designed exercises with solutions, and intuitive graphical illustrations. Readers will gain the ability to see, run, and understand causal discovery methods in practice.
Key features of this book include:
- A clear and self-contained introduction, bridging probability, statistics, and modern causal discovery techniques
- 100 exercises with solutions, supporting self-study and classroom use
- Reproducible Python code, allowing readers to implement and extend the methods themselves
- Intuitive figures and visual explanations that clarify abstract concepts
- Broad coverage of applications within statistics and data science, connecting rigorous theory with modern machine learning and causal inference
Sommario
A Gentle Introduction to Causal Discovery.- Foundations of Probability and Statistics.- Graphical Models.- Testing Independence and Conditional Independence with Kernels.- The PC Algorithm.- LiNGAM.- Information Criteria and Marginal Likelihood.- Score-Based Structure Learning.
Info autore
Joe Suzuki is a professor of statistics at Osaka University, Japan.
Riassunto
Beginning with a gentle introduction to causal discovery and the foundations of probability and statistics, this textbook is written in a highly pedagogical way. By uniting probability theory, statistical inference, and graph theory, the book offers a systematic pathway from foundational principles to cutting-edge algorithms, including independence tests, the PC algorithm, LiNGAM, information criteria, and Bayesian methods. Far more than a theoretical treatment, this volume emphasizes hands-on learning through Python implementations, carefully designed exercises with solutions, and intuitive graphical illustrations. Readers will gain the ability to see, run, and understand causal discovery methods in practice.
Key features of this book include:
- A clear and self-contained introduction, bridging probability, statistics, and modern causal discovery techniques
- 100 exercises with solutions, supporting self-study and classroom use
- Reproducible Python code, allowing readers to implement and extend the methods themselves
- Intuitive figures and visual explanations that clarify abstract concepts
- Broad coverage of applications within statistics and data science, connecting rigorous theory with modern machine learning and causal inference
