Book review: Neural Networks for Pattern Recognition By Christopher M. Bishop (Clarendon Press, 1995)

ABSTRACT

Technical textbooks tend to be encyclopedic omnibus volumes, since the market rewards authors whose books can be used by teachers in many different classes. Pedagogically, this tendency is unfortunate, since it tends to homogenize the unique approaches and presentations that make great teachers truly effective. Bishop's textbook on neural nets is a welcome exception to the rule. It specifically limits its focus to feed-forward nets with explicit training. Much important work has been done in domains such as self-organizing nets, or nets with feedback, or fuzzy algorithms, but this volume avoids them to focus on the most basic network architecture. With this sharp focus, Bishop uses several excellent pedagogical techniques.