Independent Component Analysis

Rating:
Summary: Comprehensive Book
Review: Independent Component Analysis is a young and interesting topic that gained attention and still receiving more of it.

Until now this is the best introduction that has been written.
It is comprehensive, clear and unbiased.

I think that the book is a step toward making the subject not only a common field of research but also a reference for those looking for new challenging topics.

What worths mentioning is that the authors are very envolved in the development of the theory of ICA ,other books are good but are deviated by their author's own approachs and this is normal but unhealthy for a first book on any field.

What constitutes a great help for understanding ICA are the relatively easy concepts if one just intend to pick an algorithm(ex:FastICA), but this is not the case regarding its theory.

One colleague once argued that ICA should have emerged long before the begining of the 90's, claiming that Gaussian forms
(Central Limit-Theorem) killed the idea of dealing with other kinds of distributions and therefore the signal processing community went assuming every thing was gaussian (noise was gaussian,signals are gaussian),but the emerge of HOS relaxed the gaussian restriction and ICA became possible and no longer 'blind' .

I think this should prepare researchers to deal with coming challengs more intelligently and efficiently .That is why I recommend this book since it tries to give a broad view to the subject .

Rating:
Summary: Nice and detailed description of ICA
Review: This is a nice and self-contained book on the subject of independent component analysis (ICA). The authors start with relevant mathematical and statistical background (in Part I) to prepare readers for the derivations of ICA (though seasoned researchers may want to skip the first part of this book). The authors discuss the motivation behind ICA and present several ways to derive ICA (since this subject has been approached by several communities). The authors also compare and discuss the pros and cons of these approaches. The authors discuss several applications using ICA in Part III.

Compared with other ICA books, this manuscript has much depth and completeness. I highly recommend this book to any reader interested in this topic.