on 15 May 2001
This is an excellent book which has been widely adopted around the world. The coverage is extensive and includes most of the modern techniques of Adaptive signal processing including Neural Networks etc. The organization of the text and the chapters is in a clear logical progression.
As a text book however I feel that it is perhaps too ambitious for a person new to the topic. I would have preferred lucidity of explanations over the depth of coverage for a text book of a course. My favourite book for this is still Woodrow and Wilson's classic text even though it is quite old and does not cover most new developments in the field.
However as a reference for those already somewhat aware of the topics, or for the serious DSP researcher, this book is an excellent resource which gives a very good review of most of the published material in the area.
on 10 August 2009
This review is based on the Third Edition, a copy of which I have owned for some time now. I am a practising firmware/software engineer specialising in realtime control, DSP and datacoms, with pretensions to advanced techniques. I have a good grasp of matrix maths from reading Gilbert Strang's excellent texts but Prof. Haykin's book is too light on explanation and example to give much insight without considerable recourse to other material - the Matlab Signal Processing help files are good - and progress with his book is hard going. Of course trying to master difficult subjects with one text only is not really smart but in point of fact there aren't really ANY good pedagogical expositions in Adaptive Filters (see Sayed for example). I am trying Pedro Diniz' updated text now and it does look more promising as the supporting material is well developed. I have no doubt Prof. Haykin is masterful but could he have a little pity for us lesser mortals please.