Customer Reviews

The Text Mining Handbook: Advanced Approaches in Analyzing Unstructured Data

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

 

 

This was one of the few books that included a very clear and extensive treatment of information extraction techniques. There were plenty of diagrams which is great for a visual learner. All the techniques are explained using both plain English and formulas, so that you can pick up the scientific notation with minimal previous knowledge.

Even when the authors plug their own company and research at the end it was moderately useful in illustrating the concepts mentioned in a real world scenario.

This book is a decent overview of the field but I don't think it will prepare anyone to actually create text mining systems. Feels like its oriented to technical managers rather than the people who will be actually building the stuff. Examples are really lagging, the very little psuedo code that does exists is not that good. It has very good definitions of concepts and explains what types of algorithms you'll be using but its very much like an encyclopedia. After reading through this book i quickly purchased Konchandy's Text Mining book to see if it was better. I haven't finished it yet but so far its great, would recommend that book over this one if you actually have to build text mining systems. On the up side, this book has very comprehensive references to find more information on each section which could prove very useful.

I was very disappointed by this book. It has some good material but examples are really missing. It might be a good book for technical managers to have a high level picture of the field. As a graduate student I don't seem to be able to use this book in any practical project.

Ronen Feldman and James Sanger offer an introduction to text mining and a high-level tour of its major techniques. Their goal is to provide "...a comprehensive discussion of the state of the art in text mining and link detection." The book's twelve chapters survey this field as it existed in 2006, focusing on general text mining processes. It also discusses data preprocessing, knowledge representation of processed data, and user interface design of text mining tools. The closing chapter reviews three examples of well-designed text mining applications.

I found the less central chapters to be the most useful. The third chapter on "Text Mining Preprocessing Techniques" highlighted representative "messy data" problems encountered with text and standard techniques to deal with them. I would have appreciated more detail in this chapter, but the Citations and Notes showed me where to go find it. The two chapters on methods for "browsing" knowledge extracted from text and discovering and displaying new information were first-rate. They communicated a solid understanding of the researcher's work and how well-designed software could support it. Nicely done.

The book is a reasonable presentation of basic text mining concepts and is worth reading by someone new to this field. It does NOT provide "...an in-depth examination of core text mining and link detection algorithms and operations." The depth and detail just aren't there. The good news is that a high-level description of the field ages slowly, keeping the book relevant longer. It remains worth reading as an overview.

For the missing technical detail, you might supplement with Manu Konchandy's Text Mining Application Programming from 2006 or something a little more recent. Linoff and Berry's Data Mining Techniques strikes a good compromise between overview and detail for data mining overall, but only devotes one chapter to text mining. Either is a good companion volume to this one.

The amount of textual information floating around on the web is staggering. And the overwhelming amount of this information is simply passed along from sender to receiver for the human being reading it to make sense out of it. There are obvious demands to use a computer to 'read' this material and select out appropriate nuggets from the ore.

Intended for use by advanced undergraduate students, graduate students, researchers and people working in the field, this book first covers the definition of the problem and presents several state-of-the-art probabilistic models for information extraction, and how these models can be used in applications. Finally a rather detailed description of several real life applications are included: patent searching, scanning magazine articles, and scanning the news for business intelligence.

Do you suppose that somewhere, some body (or maybe a lot of bodies) are using these techniques to find words like bomb, explosion, etc.