7 of 8 people found the following review helpful:
5.0 out of 5 stars
A great handbook for data mining, 12 April 2006
By James Franklin - Published on Amazon.com
This review is from: Data Mining and Knowledge Discovery Handbook (Hardcover)
I'm surprisingly pleased with this book. The book is well-written and it is completely worth the price. The main chapters of the book are independent, so you can read them in any order. It nearly cover the entire data mining field. In fact you can find a good overview about almost all important data mining techniques. Moreover most of the algorithms are presented in pseudo code, so you can really learn how to implement them. I also liked the application section which describes real case studies - it gives you a good sense how to use these techinques.
3 of 3 people found the following review helpful:
3.0 out of 5 stars
The Curate's Egg, 15 Jun 2008
By Chris Hobbs "cwlh" - Published on Amazon.com
This review is from: Data Mining and Knowledge Discovery Handbook (Hardcover)
This is a book (calling it a "handbook" implies that data miners have particularly large hands) of papers, loosely divided into subject areas.
The first thing to be said is that the index is rubbish. For such a large book it is totally inadequate (5 pages for a 1383 page book!), the level of indexing fluctuates wildly and there are some really strange errors:
Decision Tree 1114
Decision support systems ......
Decision table majority 99, 105
Decision tree 150, 165, 167, 314
That "Decision Tree" should be sorted away from "Decision tree" is understandable to anyone who knows ASCII but surely even then the two entries shouldn't be split by "Decision support". I also liked:
GLM (Generalized Linear Model) 240, 575
GLM (Generalized Linear Models) 213, 215
In general, the quality of the index reflects the general quality of the editing: poor. I have been involved in contributing a chapter to a book of this sort and know that it is very difficult for the editor to maintain a constant vocabulary, level and thrust throughout a book of contributed papers but this one is worse than normal: some authors simply repeat what previous authors have said, some contradict.
So, is this worth the best part of $200? If you are looking for a review of data mining techniques without a great deal of mathematical maturity being required then this is probably a reasonable book. Many of the papers cover the ground of a particular technique very well. What is lacking is the map of the wood as well as the details of each tree. There is one overview paper at the front but it introduces terminology that is not generally followed later.
The final section, examples of the use of data mining in various industrial situations (finance, telecommunications, etc.), is very superficial and does not tie into the other papers at all. It would have been worthwhile for the editor to have collected the earlier papers (tools and techniques) and provided these to the writers of the industrial examples. It would have been really useful to have tied the examples to the particular techniques: "Using the XYZ technique as described in Fred's paper on page 123 of this handbook, ....".
In summary, I don't regret having bought this book and I have learned from it but, with strong editing, it could have been 1000 times better.