Other Sources

Here are some other sources of information about the topics in the course. This includes both simple and advanced treatments of the material; the simpler texts are listed first.

• Adaptive Pattern Recognition and Neural Networks, Yoh-Han Pao (1989).
  This is a very good book. It was a textbook for the course, but is now out of print. It is good on neural networks; less good on Bayesian approaches. It contains more mathematical detail and advanced material than Weiss and Kulikowski. A very good book to get from the library and follow throughout the course.

• Neural Computing, Beale and Jackson, 1990

  An elementary book on neural networks. Good treatment of Bayesian classification, neural networks, and some of the search aspects of the course at a simple level. Contains nothing on validation or non-numerical methods.

• Machine learning, Neural and Statistical Classification, D. Michie, D.J. Spiegelhalter and C. C. Taylor, 1994

  Not a textbook. A collection of articles describing current applications of and approaches to the techniques studied in the course.

• Techniques in Computational Learning C. J. Thornton, 1992.

  A book on methods of learning from data, including supervised learning and unsupervised learning in neural networks, and learning in decision trees, as well as learning methods not covered in this course. It contains nothing on generalisation, validation, or Bayesian methods, unfortunately.

• Modern Heuristic Techniques for Combinatorial Problems, C. Reeves (editor), 1993.

  An overview of search techniques including simulated annealing and genetic algorithms.

• Genetic Algorithms in Search, Optimization, and Machine Learning, D. Goldberg, 1989.

  The standard reference on Genetic Algorithms.

• Neural Networks, A comprehensive Foundation, Simon Haykin, 1994. Second edition, 1998.

  An excellent book on neural networks. The only problem with it is that it contains almost no mention of applications. Advanced and comprehensive. Second edition is very up to date.

• Neural Networks for Pattern Recognition, Christopher Bishop, 1996

  A possible future textbook for this course -- this book contains an excellent treatment of Bayesian approaches, neural networks, and generalisation, as well as other aspects not treated here. Somewhat more advanced in its treatment, its study will give a deep understanding of the material.

• Pattern Classification and Scene Analysis, Duda and Hart, 1973

  The classic work on traditional statistical methods and Bayesian methods.

Here is the alternative reading by syllabus topic.

Bayesian Classification:

Beale and Jackson, Chapter 2; Michie, Spiegelhalter and Taylor (MST) Chapters 2,3,4; Bishop Chapters 1 and 2, Pao, Chapter 2.

Generalisation and Validation:

MST Chapter 7; Bishop Chapters 1.5, 9.1, 9.2, and 9.8.

Supervised Learning Neural Networks:

Beale and Jackson Chapters 3 and 4; MST Chapter 6; Thornton Chapters 9 - 14; Bishop Chapters 3 and 4; Pao, Chapters 5,6.

Unsupervised Learning in Neural Networks:

Beale and Jackson Chapter 5; Thornton Chapter 18.

Learning from Non-Numeric Data:

MST Chapter 5; Thornton, Chapter 5 and 6; Pao, Chapter 4.

Subsymbolic Search Techniques and Genetic Algorithms:

Reeves Chapters 1, 2, and 4; Goldberg Chapter 2.