CS 6900: Information Retrieval
Fall 2013

---

Time and Location: Mon, Wed, Fri 10:45am – 11:40am, ARC 121
Instructor: Razvan Bunescu
Office: Stocker 337
Office Hours: Mon, Wed, Fri 2:05pm – 3:00pm, or by email appointment
Email: bunescu @ ohio edu

Textbook:

Introduction to Information Retrieval by Christopher D. Manning, Prabhakar Raghavan and Hinrich Schutze. Cambridge University Press, 2008.

Supplementary Texts:

Modern Information Retrieval by Ricardo Baeza-Yates and Berthier Ribeiro-Neto. Addison-Wesley, 1999, 2011.

Course description:
This course covers the design, implementation, and evaluation of modern information retrieval systems, such as Web search engines. It will focus on the underlying retrieval models, algorithms, and system implementations, such as vector-space and probabilistic retrieval models, as well as the PageRank algorithm used by Google. The course will also cover more advanced topics in information retrieval, including document categorization and clustering, recommender systems, collaborative filtering, and personalized search.

Prerequisites:
The students are expected to be comfortable with programming and to exhibit a basic level of mathematical dexterity. Relevant background material in linear algebra and probability theory will be made available during the course.

Lecture notes:

1. Syllabus & Introduction
2. Boolean Retrieval
3. Word Distributions
4. From Text to Tokens to Terms
5. Text Processing with Python and NLTK
   • NLTK and Python examples.
6. Retrieval with Vector Space Models
7. Probabilistic Information Retrieval
8. Web Search and Web Crawling
   • Link extraction and normalization in Python.
9. Scalable, High-performance IR using Lucene
   • Lucene example for ranking with a new similarity function.
   • Lucene example for using document fields when ranking.
10. Evaluation Measures and Benchmark Datasets for IR
11. Retrieval with Language Models
    • A study of smoothing methods for language models applied to information retrieval, C. Zhai and J. Lafferty, ACM Transactions on Information Systems, 2004.
12. Link Analysis for IR
    • Most important person on English language Wikipedia? Newspaper version and Arxiv version.
      • Project: modify ranking algorithm to account for "diversity" of links/topics?
13. Web Interfaces for IR
    • Python HTTP server for a simple search engine.
14. Latent Semantic Indexing

Some of these slides are based on material from IR classes taught at UT Austin and Stanford.

Homework Assignments:

• Assignment 1
  • CACM collection.
  • Word cloud for CS course descriptions.
• Assignment 2
  • Cranfield collection.
• Assignment 3
  • OU domain queries.
  • Sample input graph and PageRank output.
  • Wiki-Large link graph.

Final Project:

• Project description
  • Apache Lucene, Nutch, and Solr.
  • Dimensionality Reduction for Information Retrieval using Vector Replacement of Rare Terms, T. Berka and M. Vajtersic, Text Mining Workshop 2011.

Other online reading materials:

• James H. Martin's Introduction to probabilities
• Jason Eisner's equestrian Introduction to probabilities
• Inderjit Dhillon's Linear Algebra Background
• Strang's Video Lectures on Linear Algebra
• Optimizing Search Engines Using Clickthrough Data, Thorsten Joachims, KDD 2002

Python and NLTK resources:

• CS 3200 lecture on Imperative programming in Python
• Python Programming Language: Official Website
• scikit-learn Machine Learning in Python
• NLTK