collaborative query previews in digital libraries
DESCRIPTION
Collaborative Query Previews in Digital Libraries. Lin Fu, Dion Goh, Schubert Foo Division of Information Studies School of Communication and Information Nanyang Technological University. Presentation Overview. Background Query Previews and Collaborative Filtering - PowerPoint PPT PresentationTRANSCRIPT
Collaborative Query Previews Collaborative Query Previews in Digital Librariesin Digital Libraries
Lin Fu, Dion Goh, Schubert FooLin Fu, Dion Goh, Schubert FooDivision of Information StudiesDivision of Information Studies
School of Communication and InformationSchool of Communication and InformationNanyang Technological UniversityNanyang Technological University
Presentation OverviewPresentation Overview
BackgroundBackground Query Previews and Collaborative Query Previews and Collaborative
FilteringFiltering Collaborative Query Previews Collaborative Query Previews
(CQPs)(CQPs) System Design and ImplementationSystem Design and Implementation Advantages of the SystemAdvantages of the System Future workFuture work
BackgroundBackgroundInformation Overload:Information Overload:
World Wide WebWorld Wide Web Digital librariesDigital libraries
Information Seeking:Information Seeking: Information seeking is a broad term encompassing the Information seeking is a broad term encompassing the
ways individuals articulate their information needs, seek, ways individuals articulate their information needs, seek, evaluate, select and use information (Lokman & evaluate, select and use information (Lokman & Stephanie, 2001) Stephanie, 2001)
Collaboration and communication are importantCollaboration and communication are important Pre-Query Information (PQI)Pre-Query Information (PQI)
Information needsInformation needs Information systemInformation system Knowledge of the collectionKnowledge of the collection
Use of PQI in Information Retrieval
Information Systems
Physical Collections
Digital Library
TargetInformation
Pre-Query Information
Information Needs
Collection Knowledge
Information Systems Query
Structure of the Collection
Domain knowledge
Example of Collection Example of Collection KnowledgeKnowledge
Suppose a user wants to search a paper on Suppose a user wants to search a paper on overview-detail style interface but does not know overview-detail style interface but does not know the title, and also a novice in this field.the title, and also a novice in this field.
The user enters “interface” or “overview, detail” The user enters “interface” or “overview, detail” as the query. However, nothing in the top 50 as the query. However, nothing in the top 50 results rings a bellresults rings a bell
Someone else searching for the same paper Someone else searching for the same paper might remember its name clearly (“Reading of might remember its name clearly (“Reading of Electronic Documents: The Usability of Linear, Electronic Documents: The Usability of Linear, Fisheye, and Overview+Detail Interfaces”). He Fisheye, and Overview+Detail Interfaces”). He knows that using “fisheye, overview, detail” as knows that using “fisheye, overview, detail” as the query keyword will yield a good resultthe query keyword will yield a good result
Concept 1: Query PreviewsConcept 1: Query Previews
Definition:Definition: Query previews provide an overview Query previews provide an overview
about the data distribution in a data about the data distribution in a data collection (Greene et al., 1999). collection (Greene et al., 1999).
Overviews are represented as aggregate Overviews are represented as aggregate information on attributes of the information on attributes of the collection---known as summary data.collection---known as summary data.
The summary data is displayed using The summary data is displayed using various visualization techniques: various visualization techniques: histograms, timelines.histograms, timelines.
Query Preview ExampleQuery Preview Example
Reduce queries with zero or large Reduce queries with zero or large number of hits. number of hits.
Prevent the retrieval of undesired Prevent the retrieval of undesired records.records.
Represent statistical information of the Represent statistical information of the database visuallydatabase visually
Advantages of Query Previews:Advantages of Query Previews:
Concept 2: Collaborative Concept 2: Collaborative FilteringFiltering
Definition:Definition: Collaborative filtering is a technique for Collaborative filtering is a technique for
recommending items to a user based on similarities recommending items to a user based on similarities between the past behavior of the user and that of between the past behavior of the user and that of likeminded people (Chun & Hong, 2001) likeminded people (Chun & Hong, 2001)
Examples:Examples: TapestryTapestry: a system that can filter information : a system that can filter information
according to other users’ annotations (Goldberg, according to other users’ annotations (Goldberg, Nichols, Oki & Terry, 1992) Nichols, Oki & Terry, 1992)
GroupLensGroupLens: a recommender system using user : a recommender system using user ratings of documents (Resnick , Courtiat & Villemur, ratings of documents (Resnick , Courtiat & Villemur, 2001)2001)
Advantages of Collaborative Advantages of Collaborative FilteringFiltering
Use the community for Use the community for knowledge sharing.knowledge sharing.
Select high quality items from a Select high quality items from a large information stream.large information stream.
LimitationsLimitations of Existing of Existing TechniquesTechniques
Query Previews:Query Previews: Lack of support for communication and Lack of support for communication and
collaboration.collaboration.
Collaborative Filtering:Collaborative Filtering: Lack of support for gathering PQI.Lack of support for gathering PQI.
Collaborative Query Previews Collaborative Query Previews (CQPs)(CQPs)
CQP is an integrated approach to augment CQP is an integrated approach to augment information seeking by supporting information seeking by supporting collaboration and communication during collaboration and communication during the process of gathering PQI.the process of gathering PQI.
CQPs generate an overview about a data CQPs generate an overview about a data collection through a set of aggregate collection through a set of aggregate information.information.
CQPs introduce a collaborative aspect by CQPs introduce a collaborative aspect by providing recommendations of queries.providing recommendations of queries.
Collaborative Query Previews Collaborative Query Previews (CQPs)(CQPs)
Direct Previews of the Data Direct Previews of the Data Collection:Collection: Through the aggregate information on selected Through the aggregate information on selected
attributes, users can get familiar with the structure of attributes, users can get familiar with the structure of the database.the database.
Recommendation of Queries:Recommendation of Queries: Through collaborative filtering techniques, CQPs Through collaborative filtering techniques, CQPs
recommend related queries previously executed by recommend related queries previously executed by other users to help the current user make better sense other users to help the current user make better sense of how the document collection met past information of how the document collection met past information needs that coincide with the present information need. needs that coincide with the present information need.
Design and ImplementationDesign and Implementation
Introduction:Introduction: ZWE ZWE provides an integrated platform for provides an integrated platform for
supporting a variety of scholarly tasks supporting a variety of scholarly tasks including browsing, querying, organizing including browsing, querying, organizing and annotating of information resources and annotating of information resources (Goh, Fu & Foo, 2002) using a spatial (Goh, Fu & Foo, 2002) using a spatial metaphor.metaphor.
ZWE supports the entire process of ZWE supports the entire process of information seeking by incorporating information seeking by incorporating CQPs.CQPs.
DesignDesign and and ImplementationImplementation
Tabs
Query previews
Artifacts (photos, metadata, annotations)
Browsingtree
Query area
Work area
Popup menu
Recommended queries
Result lists
Design and ImplementationDesign and Implementation
Multimedia Repository
Past Queries Repository
User ProfilesRepository
Searching
Browsing
Query Previews
Recommendation
Zoomable Work Environment
Authoring
Metadata Repository
Feature Extraction
Display
User Management
Design and ImplementationDesign and Implementation JAZZ: a Zoomable User Interface (ZUI) API JAZZ: a Zoomable User Interface (ZUI) API
that allows developers to quickly and that allows developers to quickly and easily build zoomable information spaces. easily build zoomable information spaces.
Design and ImplementationDesign and Implementation Tamino XML Server: Tamino XML Server: a platform to build an XML a platform to build an XML
based information retrieval system.based information retrieval system.
Database
Schema Schema Schema
XMLXML
Tamino Manager
Schema Editor
Interactive Tools
X-Query Tools
Design and ImplementationDesign and Implementation
For query recommendation module, we For query recommendation module, we proposed a hybrid approach (Fu, Goh & proposed a hybrid approach (Fu, Goh & Foo, 2003a, 2003b) to cluster past queries Foo, 2003a, 2003b) to cluster past queries and apply the algorithms to find similar and apply the algorithms to find similar past queries for a given query.past queries for a given query.
Experiments show that our hybrid Experiments show that our hybrid algorithm outperforms the existing query algorithm outperforms the existing query clustering approach. clustering approach.
Advantages of Proposed SystemAdvantages of Proposed System
Integerated work environment: more Integerated work environment: more interactive, zoomable. Multifaceted interactive, zoomable. Multifaceted information artifacts. Generic information artifacts. Generic framework. framework.
CQPs support the information seeking CQPs support the information seeking process from two perspectives: process from two perspectives: From direct previews of the data collection.From direct previews of the data collection. From queries issued previously by others.From queries issued previously by others.
Future WorkFuture Work
With the initial prototype developed, With the initial prototype developed, the next phase of this work will focus on the next phase of this work will focus on the evaluation of CQPs by users of the the evaluation of CQPs by users of the digital library. digital library.
Continuing research is also being Continuing research is also being carried out to improve the aspects of carried out to improve the aspects of query clustering by further investigating query clustering by further investigating the use of hybrid approaches, including the use of hybrid approaches, including content-based, feedback-based and content-based, feedback-based and result-based approaches. result-based approaches.
Thank YouThank You
For more informationFor more informationSchubert Foo [email protected] Foo [email protected]