 |
|
Title: Information Retrieval: Algorithms and Heuristics (The Information Retrieval Series)(2nd Edition)
ISBN: 1402030045
Author:
David A. Grossman
Ophir Frieder
Publicate Date: 2004-12-20
Publish: 2004-12-20
List Price: $59.95
Average Customer Rating: 4.0
Format: Paperback
 |
|
|
|
|
Amazon Lowest New Price: $38.11
Amazon Lowest Used Price: $29.98
Amazon Merchant Price: $42.54
|
|
|
|
|
|
|
|
|
|
|
| Customer Review: |
|
1: Easily the best intro to the IR field - Simple, very readable, practical and directly applicable
Easily the best introduction to the field of IR. Its hallmarks are:
1. very readable (concepts presented in layman terms) and directly applicable (you can literally read and apply the concepts in the real world)
2. excellent survey of the field with an comprehensive compedium of references for further reading (surveys, topical and detailed references)
3. the only book with latest information on IR strategies and utilities - so far (May 2008)
For the learnings you will get out of this book - it is a real BARGAIN. You could easily spend hours and hours of your time just trying to figure out what to read. Its a gem. Its not expensive. Buy it->implement it->realize success from it! buy it now :).
To improve your understanding from a novice level to an intermediatre level, would recommend the following books:
A. Introduction 1st book: Information Retrieval (this book) David Grossman and Ophir Frieder
B. Modern Information Retrieval by Ricardo Baeza-Yates (more technical and deeper)
C. Managing Gigabytes: Compressing and Indexing Documents and Images by by Ian H. Witten, Alistair Moffat, Timothy C. Bell (best for implementing and a decent overview)
D. Programming Collective Intelligence: Building Smart Web 2.0 Applications by Toby Segaran (algos/pseudo-code, good utilities and concise summaries)
F. Mining the web, by Chakrabarti (excellent intro the field of web mining, with some excellent chapters on IR)
G. Books by Salton (vector models and fuzzy sets)
H. IR book by C. J. Rijsbergen (probabilistic models; available online for free)
I. the book by Lesk et. al.
J. Follow-up by reading a ton of papers available on citeseer or via google
Good luck!
|
2: Poor content and presentation
This book presents information retrieval in an incomprehensible fashion. The content appears to have been cut and pasted from diverse unrelated sources with no effort put into massaging the parts into a coherent whole. There are several mistakes, typos, and wrong formulas throughout the book. Often concepts and abbreviation pop out without any context. Loose statements that are vague and hard to understand are made in several places. The book also fails to use a consistent style in its presentation. Overall, it does little to explain anything in Information Retrieval.
|
3: A good alternative to "Modern Information Retrieval"
This is a very clear and current book on information storage and retrieval. If you are assigned this book as a textbook in a class, then the book is going to make the task of understanding the material much easier. All of the algorithms are clearly explained and the background material in probability is clearly outlined with good examples and figures. However, I still think I prefer Modern Information Retrieval for the theory of information storage and retrieval. It's out of print, but you can easily find it used and just like in this book, all of the background mathematics is outlined in regards to the algorithms and tasks at hand. The other book I'd recommend is Programming Collective Intelligence: Building Smart Web 2.0 Applications. That book is on the cutting edge of using information retrieval techniques for web applications with plenty of code examples. Even if you go with this book over Modern Information Retrieval as a main source of instruction, Programming Collective Intelligence is a good book about information retrieval in combination with artificial intelligence as it is applied to the web.
|
4: Excellent coverage of IR topics
This book provides an excellent blend of theoretical and practical knowledge of the IR field, particularly for those of us with a computer science background, yet no practical working experience in IR. In my opinion, the math is an essential part of expressing the concepts more formally, so it was refreshing to see the authors incorporate just enough formulae, but no more. This book is not going to provide you with a set of recipes for building an indexing or search engine, nor would I expect it do so. However, it does give you an idea of how such engines might be built. Further, I found this book to be a necessary prerequisite for other practitioner-oriented texts, such as Lucene in Action (In Action series). Anyone delving into this field for the first time and attempting to use libraries like Lucene may find it difficult to fully exploit its capabilities without a firm understanding of the theoretical underpinnings of IR.
|
5: Mistakes in Bayes explanations
Contains a bad mathematical mistake in section 2.2.1 on page 22.
The probability P(win|sunny,good-shortstop) cannot be derived from P(win|sunny) and P(win|good-shortstop). It can take any value, even zero.
Suppose, shortstop is a vampire and plays good only on cloudy weather, and on sunny weather he always leads his team to defeat. It doesn't contradict to having positive P(win|sunny) and P(win|good-shortstop).
|
|
|
|
|
