随着人工智能、信息检索和海量数据处理等技术的发展，模式识别成为了研究热点。

在本书中，Ripley将模式识别领域中的统计方法和基于神经网络的机器学习这两个关键思想结合起来，以统计决策理论和计算学习理论为依据。建立了神经网络理论的坚实基础。在理论层面。本书强调概率与统计；在实践层面，则强调模式识别的实用方法。

本书已被国际知名大学采用为教材，对于研究模式识别和神经网络的专业人士，也是不可不读的优秀参考书。

本书是模式识别和神经网络方面的名著，讲述了模式识别所涉及的统计方法、神经网络和机器学习等分支。书的内容从介绍和例子开始，主要涵盖统计决策理论、线性判别分析、弹性判别分析、前馈神经网络、非参数方法、树结构分类、信念网、无监管方法、探寻优良的模式特性等方面的内容。

本书可作为统计与理工科研究生课程的教材，对模式识别和神经网络领域的研究人员也是极有价值的参考书。

1 Introduction and Examples

 1.1 How do neural methods differ?

 1.2 The patterm recognition task

 1.3 Overview of the remaining chapters

 1.4 Examples

 1.5 Literature

2 Statistical Decision Theory

 2.1 Bayes rules for known distributions

 2.2 Parametric models

 2.3 Logistic discrimination

 2.4 Predictive classification

 2.5 Alternative estimation procedures

 2.6 How complex a model do we need?

 2.7 Performance assessment

 2.8 Computational learning approaches

3 Linear Discriminant Analysis

 3.1 Classical linear discriminatio

 3.2 Linear discriminants via regression

 3.3 Robustness

 3.4 Shrinkage methods

 3.5 Logistic discrimination

 3.6 Linear separatio andperceptrons

4 Flexible Diseriminants

 4.1 Fitting smooth parametric functions

 4.2 Radial basis functions

 4.3 Regularization

5 Feed-forward Neural Networks

 5.1 Biological motivation

 5.2 Theory

 5.3 Learning algorithms

 5.4 Examples

 5.5 Bayesian perspectives

 5.6 Network complexity

 5.7 Approximation results

6 Non-parametric Methods

 6.1 Non-parametric estlmation of class densities

 6.2 Nearest neighbour methods

 6 3 Learning vector quantization

 6.4 Mixture representations

7 Tree-structured Classifiers

 7.1 Splitting rules

 7.2 Pruning rules

 7.3 Missing values

 7.4 Earlier approaches

 7.5 Refinements

 7.6 Relationships to neural networks

 7.7 Bayesian trees

8 Belief Networks

 8.1 Graphical models and networks

 8.2 Causal networks

 8 3 Learning the network structure

 8.4 Boltzmann machines

 8.5 Hierarchical mixtures of experts

9 Unsupervised Methods

 9.1 Projection methods

 9.2 Multidimensional scaling

 9.3 Clustering algorithms

 9.4 Self-organizing maps

10 Finding Good Pattern Features

 10.1 Bounds for the Bayes error

 10.2 Normal class distributions

 10.3 Branch-and-bound techniques

 10.4 Feature extraction

A Statistical Sidelines

 A.1 Maximum likelihood and MAP estimation

 A.2 The EM algorithm

 A.3 Markov chain Monte Carlo

 A.4 Axioms for conditional independence

 A.5 Optimization

Glossary

References

Author Index

Subject Index