SIGIR 2007 Proceedings

Demonstration

EXPOSE: Searching the Web for Expertise
Fabian Kaiser
University of Stuttgar t Universitaetsstrasse 38 70569 Stuttgar t, Germany

Holger Schwarz
University of Stuttgar t Universitaetsstrasse 38 70569 Stuttgar t, Germany

Mihály Jakob
University of Stuttgar t Universitaetsstrasse 38 70569 Stuttgar t, Germany

fabian.kaiser@ipvs.unistuttgar t.de Categories and Subject Descriptors

holger.schwarz@ipvs.unistuttgar t.de

mihaly.jakob@ipvs.unistuttgar t.de

H.3.3 [Information Search and Retrieval]: Information filtering; H.3.3 [Information Search and Retrieval]: Search process

The key concepts of EXPOSE to achieve this are as follows: · It supports users in controlling and redirecting the search instead of fully automating the process. This is crucial for precision reasons. Therefore, EXPOSE helps the user to aggregate and handle huge search results. · Enhanced exploration of the search context of promising resources and sites via focused crawling techniques. Existing approaches have been extended to meet specific concerns of identifying experts. · Intelligent integration of common search engines to gain from their knowledge about web content and meta information like structural interconnection of it's resources. Advanced search engine functionality like backlinks or the search for similar resources is utilized. · Calculation and visualization of the conceptual context of web resources or queries and the use of this context in subsequent searches. · Identification of persons, exploration and visualization of their interconnections with other persons as well as with the resource context. · Expertise ranking of identified persons based on several criteria like how often they were mentioned, the relevance of the resources they appear on, the textual context in which they appear, their interconnection with other persons and resources etc. In this demonstration we will present EXPOSE, our solution to finding experts on the web. We show how EXPOSE supports the user in diverse tasks throughout the whole search process and how using EXPOSE can improve the result quality compared to ad-hoc searches with common web search engines.

General Terms
Algorithms, Design

Keywords
Expert Finding, Search Engine, Information Retrieval, Web Search, Knowledge Management

1. EXTENDED ABSTRACT
Knowledge is the key factor to economic success. While more and more knowledge is explicitly written down and freely available to anybody, there is still a lot of knowledge and expertise available only implicitly - in the form of experts that can be consulted. In practice, however, a common problem is that for a given topic, these experts are unknown. Companies have been trying to solve this problem for years by running in-house expert finder tools, yellow pages systems and several other knowledge management tools. However, especially in smaller companies, the required knowledge often is simply not available in-house due to missing human resources and thus the required experts cannot be found using the mentioned approaches. To widen the search space and thus the number of potential experts that can be found, we developed the Expert Search Engine EXPOSE [1] that supports users in finding experts for arbitrary topics on the web. Our approach utilizes the fact that with the WWW, a huge data resource is available that can be searched for information about people or companies that are likely to have expertise on the topic in question. However, in contrast to enterprise knowledge management systems, a web based expert finder tool cannot rely on information about content structures as typically available in intranet scenarios. Instead, it has to cope with unstructured information resources. The definition of heuristics how to access this unstructured data and extract structured information as well as to decide to what extent a person matches the given search criteria are ma jor issues in this context.
Copyright is held by the author/owner(s). SIGIR'07, July 23­27, 2007, Amsterdam, The Netherlands. ACM 978-1-59593-597-7/07/0007.

2. ACKNOWLEDGMENTS
This work was conducted in the context of the pro ject novanet, which is funded by the German Federal Ministry of Education and Research (BMBF).

3. REFERENCES
[1] F. Kaiser, H. Schwarz, and M. Jakob. Finding experts on the web. In J. A. M. Cordeiro, V. Pedrosa, B. Encarna¸~o, and J. Filipe, editors, WEBIST (1), ca pages 363­368. INSTICC Press, 2006.

906