On Information Need and Categorizing Search

This dissertation is located in the area of information retrieval (IR), a discipline that deals with the task of satisfying a person’s information need with the help of a computer. IR systems let a user specify an information need, which is evaluated against large collections of digital documents. Techniques for satisfying information needs have meanwhile become ubiquitous, be it in the form of search engines at home or at work, on mobile devices or on workstations, or as retrieval components in file systems, document repositories, databases, or knowledge management tools. The thesis develops new concepts and algorithms in various aspects throughout the information retrieval process, with a special focus on automatic document categorization.