A Background in Signal Processing
Haykin comes to this book from an extensive background in the theory of
digital signal processing, particularly in adaptive filters. It can almost
be said that he treats neural networks as a type of digital filter.
In many ways, this is quite appropriate. However, I also believe that the
approach suffers somewhat in that the topics which are stressed are often not
those in current use by the practitioners of neural network applications.
The coverage, however, is thorough in those areas which are covered.
The technique is statement, statement and restatement, with a strong
emphasis on the mathematical nature of the topic.