Information retrieval

Course Code :2001WETINR
Study domain:Computer Science
Academic year:2019-2020
Semester:1st semester
Contact hours:60
Credits:6
Study load (hours):168
Contract restrictions: No contract restriction
Language of instruction:English
Exam period:exam in the 1st semester
Lecturer(s)Toon Calders

3. Course contents *

The course covers the foundations of current ranking techniques (which are often based on statistical models), index structures, and current implementation issues for the design of effective and scalable information-system architectures. The course covers a wide range of ranking principles, starting from the Boolean retrieval model, over to statistical ranking models such as TF-IDF, and on to probabilistic ranking techniques such as Okapi-BM25. The course will cover classical IR topics, including link-analysis methods such as PageRank, the user-specific personalization of queries, and relevance feedback. 

The course topics are:

 

  • Boolean retrieval
  • Vector Space Model
  • Language Based Model
  • Boolean Independence Model
  • Top-K querying and Index construction
  • Index construction and compression
  • Feedback-Expansion
  • Evaluating IR
  • Link analysis
  • Dimensionality Reduction
  • Neighbor search in high dimensional data