Customer Reviews

Encyclopedia of Data Warehousing and Mining

5 star:		(0)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

I'm not entirely sure why I'm bothering to write this review, since at $500, only corporate and academic libraries can afford to buy this set of books.

Well, anyway, it's hard to rate a book like this. It's not really an encyclopedia, despite the title. It's more like a crash-course on the state-of-the-art in a variety of fields. If you are a graduate student or an engineer who rapidly needs to find out what's going on in the field of, say, association rule learning or fuzzy clustering or unsupervised discretization or microarray mining, this is a good place to start. You can probably find an article here on your topic of interest, and that article will probably have references to the recent important papers in the field (although the references may be skewed toward publications by the article's author and his colleagues.) I obtained these books from my university's library and I found them to be a quick way to track down a bunch of recent papers on the topics I was interested in. However, beyond providing these references, I can't say that the actual articles in the collection are very helpful... they are generally very short, and in the few cases where they actually contain any substance, it is often a condensed version of the author's own work. (There are some exceptions.)

With regard to the design of the book(s), they could have done a lot better. The consistency of the articles varies pretty widely (which probably can't be helped), but the type-setting and editing also leave something to be desired; in many cases, equations and graphics do not look professionally typeset. The organization of the articles is perplexing at best. The articles are arranged alphabetically by title, but the titles do not seem to have undergone any editorial review. Therefore we end up with "Microarray Data Mining" on p.728 and "Mining Microarray Data" on p.810, separated by 14 other articles. There are a large number of other examples like this, as well as a substantial number of titles that begin with "Data Mining..." or "Mining...," which makes the alphabetical organization of the book fairly unhelpful. There are no cross-references between articles, and the index is also very weak, since it only indexes words that appear in the "key terms" section of each article. Therefore, in order to find all the articles on your topic of interest, your only real recourse is to scan through the entire table of contents. (It seems that in our modern era of fully-searchable online documents, people are forgetting some important aspects of making usable printed reference books.)

Finally, the copy that I obtained (which was probably just recently purchased by the library) has one page which is not at all attached to the binding. It was not ripped out -- it was just never attached. I have never seen this with any other book at any price (let alone in the $500 range!). I don't know what kind of binding operation they used on this, but I think that a $500 book set should come with all the pages attached.