Information Retrieval

Lecturer: Dr K. Spärck Jones (ksj@cl.cam.ac.uk)

No. of lectures: 8

Prerequisite courses: none, although a basic encounter with Probability is assumed

Aims

The course aims to characterise information retrieval in terms of the data, problems and concepts involved, and the formal models on which systems are grounded. These will be illustrated by example systems and performance analysis.

Lectures

• Basic problems. Capturing information content so that documents relevant to a user's information need can be retrieved. Examples.

• Key concepts and facts of life. Indexing and indexing language, search strategy and matching function, and constraints on these. Examples.

• Formal models 1. Ground abstractions, quantitative models for retrieval systems: vector model, basic concepts: treatment of indexing and matching.

• Formal models 2. Probabilistic model, basic concepts: treatment of indexing and matching.

• Implementation techniques and systems. Term selection and weighting, term and document clustering, matching functions and relevance feedback. SMART and OKAPI systems.

• Evaluation issues. Performance criteria and effectiveness measures, test methodology, established results.

• Advanced techniques and systems. Natural language processing, multiple key types and complex matching. NYU/GE and INQUERY systems.

• Present needs and opportunities. Finding pins in haystacks, reaching multimedia dreamworlds, delegating information seeking and extraction tasks.

Objectives

At the end of this course, students should be able to

• differentiate document retrieval from neighbouring areas, for example conventional database management

• appreciate the challenges of document retrieval, especially from very large files

• understand and develop concepts and strategies suited to specific retrieval situations

• apply and test chosen approaches in system implementations

Recommended books

Sparck Jones, K. & Willett, P. (eds.) (1997). Readings in Information Retrieval. San Francisco: Morgan Kaufmann.
(This is a substantial collection of key papers, along with introductory material both for the whole volume and individual sections that acts as a mini text book.)

Baeza-Yates, R. & Ribiero-Neto, B. (1999). Modern Information Retrieval. Reading, MA: Addison-Wesley and ACM Press.
(A solid wide-coverage textbook, including current multimedia, WWW etc. Much material on mechanics.)

NB This book is much better than...

Frakes, W.B. & Baeza-Yates, R. (eds.) (1992). Information Retrieval: Data Structures and Algorithms. Englewood Cliffs NJ: Prentice-Hall.)

Willett, P. (ed.) (1988). Document Retrieval Systems. London: Taylor Graham.
(A useful collection illustrating key ideas underlying modern systems.)

van Rijsbergen, C.J. (1979). Information Retrieval. London: Butterworths.
Available online at http://www.dcs.gla.ac.uk/Keith/Preface.html
(Good and clear on basic theory, but naturally lacks cover on some of their recent development.)

Salton, G. & McGill, M.. (1983). Introduction to Modern information Retrieval. New York: McGraw Hill.
(Useful material, but rather dense.)