概率图模型是一类用图形模式表达基于概率相关关系的模型的总称。概率图理论共分为三个部分，分别为表示理论，推理理论和学习理论。目前在图像和视频智能信息处理领域已有应用，基本的概率图模型包括贝叶斯网络、马尔可夫网络和隐马尔可夫网络。本文是一篇很好的图模型入门资料。 The concept of modelling using graph theory has its origin in several scientific areas, notably statistics, physics, genetics, and engineering. The use of graphical models in applied statistics has increased considerably over recent years and the theory has been greatly developed and extended. This book provides a self-contained introduction to the learning of graphical models from data, and is the first to include detailed coverage of possibilistic networks - a relatively new reasoning tool that allows the user to infer results from problems with imprecise data. One major advantage of graphical modelling is that specialised techniques that have been developed in one field can be transferred into others easily. The methods described here are applied in a number of industries, including a recent quality testing programme at a major car manufacturer., * Provides a self-contained introduction to learning relational, probabilistic and possibilistic networks from data, * Each concept is carefully explained and illustrated by examples, * Contains all necessary background, including modeling under uncertainty, decomposition of distributions, and graphical representation of decompositions, * Features applications of learning graphical models from data, and problems for further research, * Includes a comprehensive bibliography, An essential reference for graduate students of graphical modelling, applied statistics, computer science and engineering, as well as researchers and practitioners who use graphical models in their work. is the first to include detailed coverage of possibilistic networks - a relatively new reasoning tool that allows the user to infer results from problems with imprecise data. One major advantage of graphical modelling is that specialised techniques that have been developed in one field can be transferred into others easily. The methods described here are applied in a number of industries, including a recent quality testing programme at a major car manufacturer., * Provides a self-contained introduction to learning relational, probabilistic and possibilistic networks from data, * Each concept is carefully explained and illustrated by examples, * Contains all necessary background, including modeling under uncertainty, decomposition of distributions, and graphical representation of decompositions, * Features applications of learning graphical models from data, and problems for further research, * Includes a comprehensive bibliography, An essential reference for graduate students of graphical modelling, applied statistics, computer science and engineering, as well as researchers and practitioners who use graphical models in their work.