COSC 488 Syllabus Dr. Nazli Goharian nazli@ir.cs.georgetown.edu - - PDF document

1

COSC 488 Syllabus

• Dr. Nazli Goharian

nazli@ir.cs.georgetown.edu http://cs.georgetown.edu/~nazli

Office: 333 St. Mary’s Hall Tuesdays: 11-12 Thursdays 11-12 OR by appointment!

Course Description

Overview of fundamental issues of information retrieval with theoretical foundations. The Information-retrieval techniques and theory, covering both effectiveness and efficiency of information-retrieval systems are covered. The focus is on algorithms and heuristics used to find documents relevant to the user request and to find them fast. The course covers the architecture and components of the search engine such as document pre-processing, index construction, and query processing. The students learn the material by building a prototype of such a search engine.

2

Prerequisite

COSC 160 (Data Structure) & VERY STRONG Programming knowledge

The class is VERY intensive in projects. If you are not VERY comfortable in programming, you may find this course overly difficult.

Recommended Textbooks

Any one is fine (must have & read!):

• C. Manning, P. Raghavan, H. Schutze, Introduction

to Information Retrieval, Cambridge Publisher, 2008, ISBN: 978-0-521-86571-5

• D. Grossman and O. Frieder, Information Retrieval:

Algorithms and Heuristics, Second Edition 2004, Springer Publishers, ISBN 1-4020-3004-5 (paperback).

3

Handouts

The course handouts are available on the Blackboard for most topics that are covered in the class.

Projects: 40%

• Building a Search Engine
• Divided into 3-4 parts

– Incrementally building the engine

4

Projects (cont’d)

The projects require design and implementation of various components of a search engine per the assignment requirements, performing experimentations, and analysis. Deliverables for each project part include (detail will be specified when project is given):

– Cover Page – Design document – Software – Results & Analysis – Demo (failing to provide answers to questions asked during the demo on what you have delivered, results in a zero for that project)

Research Paper Presentation: 10%

Presentations must be extremely well rehearsed – failure to properly prepare for the presentation will result in a zero, or in the best case, an extremely poor grade on the presentation. All students must attend all sessions of the presentations and may be asked to evaluate each

• ther's presentations. Failing to do so, the student

will loose points!

5

Exams: 50%

2- 3 Exams -- TBD Exam Dates: TBD - will be announced in the class and on the Blackboard as the semester progresses. No makeup exams will be given without a medical excuse or prearrangement for an approved reason.

Late Assignment Policy

Assignments MUST be submitted on or before their due date. No late assignment will be assigned a grade .In fact the grade for a late submission will be a zero! If you are not able to finish your assignment by the due date and time, simply submit whatever of the assignment you have done to get some points rather than a zero. The students are encouraged to re-work on the incomplete assignments within a week from the submission due. This does not change the grade for that assignment, however will be considered if the final grade is in border-line. Note that the re-submission must be a complete work of that project part. Do not forget that this is your responsibility to make your submission on time!

6

Academic Integrity

Visit the Honor System Website at: http://gervaseprograms.georgetown.edu/honor/

Outline

• Introduction
• IR architecture
• Retrieval strategies (models)
• Retrieval techniques to improve effectiveness
• Evaluation
• Indexing
• Efficiency in indexing and query processing
• Integrating structured data and text
• Distributed IR: Web Retrieval
• Text Classification
• Recommendation systems