Misc3004: Unterschied zwischen den Versionen

Version vom 13. Dezember 2010, 22:13 Uhr

Faceted Semantic Search

Veröffentlichung: 2010 Januar
Art der Veröffentlichung: Technical Report
BibTeX

Kurzfassung
Faceted search allows users to browse and discover relevant items from a large collection such as the Web. Given a large amount of facets, current solutions allow users to focus on a ranked list of facets. While state-of-the-art ranking mechanisms are either generic or assume that users are searching for specific items, the solution we propose targets fuzzy information needs. We want to support users in discovering new or unfamiliar items of interest. Thus, instead of supporting users in searching for specific items, we propose a ranking scheme that is oriented towards browsing and exploration. Also targeting effective browsing, we propose clustering and facet grouping mechanisms to allow users to uniformly browse facets and facet values. Via a task-based evaluation, we demonstrate that the proposed solution enables more effective browsing, when compared to the state-of-the-art that is focused on precise needs.

Download: Media:Paper-tr.pdf

Forschungsgruppe

Wissensmanagement

Forschungsgebiet

Information Retrieval, Semantische Suche

@@ Zeile 10: / Zeile 10: @@
 }}
 {{Publikation Details
-|Abstract=The increasing amount of data on the Web bears potential for addressing complex
+|Abstract=Faceted search allows users to browse and discover relevant items from a large collection such as the Web. Given a large amount of facets, current solutions allow users to focus on a ranked list of facets. While state-of-the-art ranking mechanisms are either generic or assume that users are searching for specific items, the solution we propose targets fuzzy information needs. We want to support users in discovering new or unfamiliar items of interest. Thus, instead of supporting users in searching for specific items, we propose a ranking scheme that is oriented towards browsing and exploration. Also targeting effective browsing, we
-information needs more eﬀectively.  Instead of keyword search and browsing along
+propose clustering and facet grouping mechanisms to allow users to uniformly browse facets and facet values. Via a task-based evaluation, we demonstrate that the proposed solution enables more effective browsing, when compared to the state-of-the-art that is focused on precise needs.
-links between results, users can specify their needs in terms of complex queries and
+|Download=Paper-tr.pdf
-obtain precise answers right away.  However, browsing is also essential on the Web
-of data as users might not always know a speciﬁc query language and more im-
-portantly, might not know the data.  Particularly in cases where the information
-need is fuzzy, browsing is useful for exploring the data. Faceted search allows users
-to browse along facets.  However, work on faceted search so far has been focused
-on search rather than browsing.  In this paper, we propose a facet ranking scheme
-that targets the browsing experience.  When there are too many facets given, user
-obtain a ranked list of facets, where the rank represents the facets’ browse-ability.
-Furthermore, facets might be associated with a large amount of values. Also target-
-ing browse-ability, we propose clustering mechanisms to decompose such facets into
-more ﬁne-grained sub-facets. By means of a task-based evaluation, we demonstrate
-that the proposed solution enables more eﬀective browsing, when compared to the
-state of the art that is rather focused on search-ability.
-|Download=FacetedSemanticSearchTR.pdf
 |Forschungsgruppe=Wissensmanagement
 }}