Modern Information Retrieval (Acm Press Series) (英語) ペーパーバック – 1999/5/15

This is an excellent book for those interested in getting an overview of IR. The book summarizes all the important milestones of IR up to 1999 (There are 852 references in the bibliography!). The writing is concise yet eloquent. The authors try to cover as much ground as possible, providing a gold-mine of information comparing the pros and cons of the various types of implementation. However, I believe that due to the breadth of the techniques covered, some of the explanations for the algorithms were rather brief and not very illuminating. But no worries, there are ample references to point you back to the writings of the orignal authors so you can get right back on track.

A very good approach to IR from a computer science perspective. Well organized and easy to read although some chapter could be dry for a novice. Excellent cover of visualization techniques, digital libraries, and web searching from a conceptual point of view. The reference list is large enough (and up-to-date) to keep you busy reading more material for several weeks.

I used this book as a textbook in a course on information storage and retrieval that I took a few years back, and it is still my favorite book on the subject. It explains the concepts clearly yet has all of the necessary mathematical and algorithmic details needed to work with the subject matter.
Chapter one just acts as a guide to the rest of the book. The book is basically divided into four parts: text IR, human-computer interfacing for IR, multimedia IR, and applications of IR. The part on text IR is best for beginners trying to learn the overall subject of IR, and consists of chapters 2 through 9. Chapter 2 is a long and important chapter that introduces fundamental concepts in IR and lays foundations for later chapters. Models for "ranking" documents based on queries are presented, including the boolean, vector, probabilistic, and fuzzy models. Chapter 3 is far less technical than chapter 2 and focuses on evaluation of IR models. Chapter 4 is an introduction to query languages, which are necessary for the elegant presentation of complex queries. Chapter 5 deals with query operations, which is the transformation of queries from simple keywords into weighted sets of terms and also includes user feedback. As in previous chapters, there is quite a bit of mathematics involved. Chapter 6 is devoted to text languages such as HTML and SGML since the user might refer to the structure of a document in his/her query, and that structure must be defined somewhere. Chapter 7 is about operations on documents themselves for the purpose of simplifying them for quick search. Thus, it is important as a time saver to eliminate common words such as "the" and also to reduce words to their grammatical roots. The potentially large size of document collections requires special indexing techniques for efficient retrieval. This is the subject of Chapter 8. Query processing can be further accelerated by using the parallel and distributed IR techniques discussed in Chapter 9, which concludes the book's discussion of text IR.
Chapter 10 is a stand-alone chapter on HCI for IR that discusses the design of user interfaces that assist the user in forming a query and current approaches for visualization of large data sets.
Multimedia IR is discussed in chapters 11 and 12. Models and query languages for office and medical information systems are discussed in Chapter 11. Efficient indexing and searching of multimedia objects is discussed in Chapter 12.
The final three chapters of the book are about the applications of IR. There is a chapter each about searching the web, bibliographic systems, and digital libraries.
The chapter on text languages is starting to show its age, as are the chapters on IR applications at the end of the book. The chapters on algorithms, and particularly the algorithmic portions of the chapters on text IR cause this book to remain a worthwhile read. There is quite a bit of mathematics used in this book, and probability theory in particular. Thus, the reader should already be familiar with probability theory and the basics of pattern recognition to get the most from this book.

A fantastic, in depth, survey of all the issues surrounding IR, from algorithms to presentation of IR results. With one clear authorial voice, the authors present all the things you hope a survey book will- a structured, coherent and complete framework onto which you can append future learning; what common practice within commercial industry really is; a quantitative analysis of the relative effectiveness of each algorithm, including the methodolgy used to arrive at results; an in-depth and clear explanation of all major algorithms.

They also give fair warning when they are only covering the outline of subject matter (which is rare), and they give extensive footnotes for anyone who needs to go deeper. The writing is always clear; the auithors never engage in the type of handwaving that other authors use to get past material you have the impression they themselves don't fully grasp.

If you need to implement search for a database and don't know where to start or what might be involved, this is the book for you. If you need to implement the GUI for search results and are wondering what the state of the art is and what issues are involved, then this is the book for you. If you need a well-structured framework to help you understand how internet search engines work, then this is the book for you. If you want to press the research forward on any of these topics and you are not already fluent in the literature, then this is the book for you.