Predictive Data Mining: a practical guide (The Morgan Kaufmann Series in Data Management Systems) (Paperback)

by Nitin Indurkhya (Author) "Data mining is the search for valuable information in large volumes of data..." (more)
Key Phrases: standard spreadsheet form, dynamic feature selection, predictive data mining, Data Preparation Raw, Data Reduction Feature, Overall Assessment Timing (more...)

Editorial Reviews

Amazon.com Review
Data mining is a hot technology, and this short, authoritative guide shows how it works and why it is gaining ground in the worlds of finance, manufacturing, marketing, and health care. The book begins by exploring the links between "big data"--the data warehouse built up of multiple databases--and traditional statistics. (The authors defend the methods of big data against traditional statistics, which has usually relied on smaller samples. However, they also look at the sources of error in both disciplines.)

The authors then look at the nuts and bolts of the data-mining process. They show how data must be prepared--sometimes reduced--in order to be manageable, and they define the important features. They show how the actual analysis of data mining can be as simple as adding up scores for selected features or how it can use statistical methods or even neural networks. (For some problems, the features themselves aren't known ahead of time; data mining can be used to discover these features automatically.) The authors then discuss how to interpret the results of analysis so that predictions can be made for new cases based on old ones.

The book concludes with short scenarios of how data mining can be applied, with examples drawn from manufacturing, health care, marketing, and publishing. The authors show the strengths--and limits--of data mining and argue that faster hardware and greater database storage capabilities will make this technology more widely used. Though it is written by two researchers in the field, Predictive Data Mining is suitable for general readers who are interested in the topic. --Richard V. Dragan

Review
"I enjoy reading PREDICTIVE DATA MINING. It presents an excellent
perspective on the theory and practice of data mining. It can help
educate statisticians to build alliances between statisticians and
data miners."
—Emanuel Parzen, Distinguished Professor of Statistics, Texas A&M University -- Review

See all Editorial Reviews

Product Details

• Paperback: 228 pages
• Publisher: Morgan Kaufmann; 1 edition (August 15, 1997)
• Language: English
• ISBN-10: 1558604030
• ISBN-13: 978-1558604032
• Product Dimensions: 8.9 x 6 x 0.6 inches
• Shipping Weight: 14.1 ounces (View shipping rates and policies)
• Average Customer Review: 4.0 out of 5 stars See all reviews (8 customer reviews)
• Amazon.com Sales Rank: #906,370 in Books (See Bestsellers in Books)

Customer Reviews

Most Helpful Customer Reviews

18 of 20 people found the following review helpful:

2.0 out of 5 stars Not that practical!, November 6, 1998

By	[e-mail] (Santiago, Chile) - See all my reviews

If you are new to data mining, you probably won't grab a global view of it with this book. According to the authors, it is not the objective, they want it to be a practical guide. Well, it's not!

On the other hand, if you know about modelling, predictive methods and try to build data mining solutions in your work environment, you will probably feel that the book is kind of useless. The hot hints they are supposed to give you are pretty straightforward and will be better learned by experience on a database.

Nothing of what I read in this book was rated 'Waow'!

I was disapointed.

7 of 7 people found the following review helpful:

4.0 out of 5 stars Reasonably good introduction, May 24, 1998

By A Customer

This isn't a bad book to pick up if you want to find out what data mining is about. I did, and it served as a good introduction.

For those of you who, like me, don't know what this is about, let me try to summarize. For years, organizations have been collecting a lot of information, via computer, just to run their business. For legal and business reasons, they have had to hold on to it, long past what they considered to be its useful life. But other than just storing it, what good is it?

Well, someone decided it could be used to answer questions about the business. Enter the data warehouse. The idea is to take all this old data, clean it up, and put it in a large database. Then the data can be mined for information.

There are two functions of data mining. One is to answer questions about the business. The other is to discover new knowledge about the business that you did not even have the sense to put in the form of a question. Everything from simple statistics to neural nets, genetic algorithms, and evolutionary programs can be used to mine the data.

Like any other science, this can be used for good purposes (what's the main reason homeless people become homeless), or bad (who's most likely to buy the Brooklyn Bridge). It can be used in many areas of science, although I suspect it will mostly be used by businesses trying to take marketing where no man has gone before.

The book itself is mostly prose, so it's an easy read, although it does require some computer knowledge. The more technical sections (like k-means and entropy clustering) are awkwardly written. But this does not detract from the overall effectiveness of the book.

If you're a manager whose boss just told you to head up the data warehouse, and you don't have a clue what he's talking about, this wouldn't be a bad book to get.

I give it a 7. It's easy to dance to.

5 of 5 people found the following review helpful:

5.0 out of 5 stars Excellent introduction, April 14, 1999

By	Amrit B. Tiwana (Atlanta, Georgia) - See all my reviews

I felt that this was an excellent intorduction to an area otherwise overcrowded with untested and semi validated "methodologies". I loved the fact that this title is not vendor specific (unlike some other titles trying to sell you tools or what!). Seeing the MKP name on it is reassuring. Would'nt hesitate to recommend this one!