INFM 700: Session 7 Unstructured Information (Part II)The IR Black BoxThe Role of InterfacesToday’s TopicsSource Selection: GoogleSource Selection: AskSource ReselectionThe Search BoxAdvanced Search: FacetsFilter/Flow Query FormulationDirect Manipulation QueriesResult PresentationAlternative DesignsBinocularsTileBarsTechniqueExampleTileBars ScreenshotTileBars SummaryScrollbar-TilebarCat-a-ConeCat-a-Cone InterfaceCat-a-Cone ArchitectureClustering Search ResultsVector Space ModelSimilarity MetricComponents of SimilarityText ClusteringThe Cluster HypothesisVisualizing ClustersTwo StrategiesHACSlide 33What’s going on geometrically?Cluster SimilarityDifferent Similarity FunctionsNon-Hierarchical ClusteringK-MeansK-Means AlgorithmK-Means Clustering ExampleK-Means: DiscussionWhy cluster for IR?From Clusters to CentroidsClustering the CollectionClustering the ResultsScatter/GatherScatter/Gather ExampleSlide 48Slide 49Slide 50Clustering Result SetsNavigation SupportPadPrintsPadPrints ScreenshotPadPrints ThumbnailsZoomable HistoryDoes it work?Slide 58INFM 700: Session 7Unstructured Information (Part II)Jimmy LinThe iSchoolUniversity of MarylandMonday, March 10, 2008This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United StatesSee http://creativecommons.org/licenses/by-nc-sa/3.0/us/ for detailsiSchoolThe IR Black BoxSearchQueryRanked ListiSchoolThe Role of InterfacesSourceSelectionSearchQuerySelectionRanked ListExaminationDocumentsDeliveryDocumentsQueryFormulationResourcesource reselectionSystem discoveryVocabulary discoveryConcept discoveryDocument discoveryHelp users decide where to startHelp users formulate queriesHelp users make sense of results and navigate the information spaceiSchoolToday’s TopicsSource selectionWhat should I search?Query formulationWhat should my query be?Result presentationWhat are the search results?Browsing supportHow do I make sense of all these results?Navigation supportWhere am I?Source SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolSource Selection: GoogleSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolSource Selection: AskSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolSource ReselectionSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolThe Search BoxSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolAdvanced Search: FacetsSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolFilter/Flow Query FormulationDegi Young and Ben Shneiderman. (1993) A Graphical Filter/Flow Representation of Boolean Queries: A Prototype Implementation and Evaluation. JASIS, 44(6):327-339.Source SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolDirect Manipulation QueriesSteve Jones. (1998) Graphical Query Specification and Dynamic Result Previews for a Digital Library. Proceedings of UIST 1998.Source SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolResult PresentationHow should the system present search results to the user?The interface should:Provide hints about the roles terms play within the result set and within the collectionProvide hints about the relationship between termsShow explicitly why documents are retrieved in response to the queryCompactly summarize the result setSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolAlternative DesignsOne-dimensional listsContent: title, source, date, summary, ratings, ...Order: retrieval score, date, alphabetic, ...Size: scrolling, specified number, score thresholdMore sophisticated multi-dimensional displaysSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolBinocularsSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolTileBarsGraphical representation of term distribution and overlap in search resultsSimultaneously Indicate:Relative document lengthQuery term frequenciesQuery term distributionsQuery term overlapMarti Hearst (1995) TileBars: A Visualization of Term Distribution Information in Full Text Information Access. Proceedings of SIGCHI 1995.Source SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolTechniqueRelative length of documentSearch term 1Search term 2Blocks indicate “chunks” of text, such as paragraphsBlocks are darkened according to the frequency of the term in the documentSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolMainly about both DBMS and reliabilityMainly about DBMS, discusses reliabilityMainly about, say, banking, with a subtopic discussion on DBMS/ReliabilityMainly about high-tech layoffsExampleTopic: reliability of DBMS (database systems)Query terms: DBMS, reliabilityDBMSreliabilityDBMSreliabilityDBMSreliabilityDBMSreliabilitySource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolTileBars ScreenshotSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolTileBars SummaryCompact, graphical representation of term distribution in search resultsSimultaneously display term frequency, distribution, overlap, and doc lengthHowever, does not provide the context in which query terms are usedDo they help?Users intuitively understand themLack of context sometimes causes problems in disambiguationSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolScrollbar-TilebarFrom U. MassSource SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolCat-a-ConeKey Ideas:Separate documents from category labelsShow both simultaneouslyLink the two for iterative feedbackIntegrate searching and browsingDistinguish between:Searching for documentsSearching for categoriesMarti A. Hearst and Chandu Karadi. (1997) Cat-a-Cone: An Interactive Interface for Specifying Searches and Viewing Retrieval Results using a Large Category Hierarchy. SIGIR 1997.Source SelectionQuery FormulationResult PresentationBrowsing SupportNavigation SupportiSchoolCat-a-Cone InterfaceSource SelectionQuery FormulationResult
View Full Document