<< 1 >>
Rating: Summary: Fascinating but the price isn't right... Review: SOM are a kind of neural networks that are totally different from the feedforward multi-layered networks that everybody has heard about. The ideas behind SOM are remarkable intellectual stuff, first intellectual class stuff, and would certainly deserve to be more widely known. I found myself wondering how comes I rarely heard of Prof. Kohonen, and when is he going to win the Nobel prize... Prof. Kohonen and the development of SOM would make a good subject for the James Gleicks out there. This book is in fact a collection of short articles about how SOM can be applied to financial visualisation and classification problems. Compilations of articles are a awkward genre, ranging from either the totally unitelligible or the totally trivial, and if you have graduate education in finance, you will find that most of the articles are pretty lame, if not downright amateurish, compared to standard academic finance literature; if you don't have grad education, I'm not sure whether you would want to buy this book in the first place, but there's nothing particularly hard about it. Art buffs may be attracted by the richly colored SOM maps (Vasarely-like) that come inside the book, and I could imagine somebody exposing these in an art gallery... Anyway, after going through this book, one gets a decent idea of what can be done with SOM, so I would definitely say IT IS worth reading. I have to say I found it utterly fascinating, although I was disappointed with most of the articles. Another thing I don't like about the book is the price (I mean, OUCH!)... The other thing is the fact that even after reading the theory article by Professor Kohonen in this book, I still couldn't understand intuitively how the basic SOM convergence algorithm worked. It is only after reading the first few pages of his masterpiece "Self-Organizing Maps", published at Springer in 1992 that I had the "haha!" experience and that I decided to put Professor Kohonen on the pedestal of my personal heroes of science...
<< 1 >>
|