Yuri de Lugt Collexis Karin Clavel TU Delft Library
Sunday, August 16, 2015 The Collexis Company Collexis develops and implements software for making large amounts of (un)structured data easily accessible Founded in 1999, 40 employees Based in USA, NL, GE –sales/development in the Netherlands (Geldermalsen) Worldwide coverage through partnerships Collexis grants free licenses to selected projects in developing countries
Customer in Library domain
Introduction The main question: “Is semantic & concept search a competitive Edge?” “The role of the expert, trivial?” Collexis introduced Case: University Library of Wageningen Case: University Library of Delft Sunday, August 16, 2015
Search Engine vs. Knowledge Engine
Sunday, August 16, 2015 Elementary principle: Validation Validation of: –Content –Source –Concept –Meaning and of interpretation How….? By expert involvement !! –Domain restricted –Targeted Content sources
Sunday, August 16, 2015 The Thesaurus – The Expert A thesaurus defines the world that we are looking at Domain experts’ expertise is used to create a thesaurus: therefore a thesaurus is validated knowledge Every user benefits from the knowledge the expert added to the thesaurus Natural language is very complex; a thesaurus helps us to ‘understand’ the natural language Find, purchase or build => validated by experts
Sunday, August 16, 2015 Levels of Ambition document-based information retrieval identify relevant terms in documents/query aggregation/clustering combine information per subset of documents association link information in a document collection Ease of use, Ease of Search, Validated results identify relevant terms in documents/query Metadata aggregation combine information per meta data and subset of documents Knowledge Discovery Explore beyond existing knowledge 1 2 3
Added value for Library Collexis exactly knows what text is all about –including homonyms, synonyms, multi-lingual aspects, Hierarchy knowledge Search and retrieval made easy –Multiple search phrases, Combining of sources, Concept- based search, Classification/keyword tagging Using the existing information –Finding experts, Knowledge actively used within the curriculum, Detection of plagiarism Sunday, August 16, 2015
Case: Wageningen UR Library Sources : –Article and Experts Search on: –Content and Metadata Ease of use, Ease of Search Deep linking to: –Repository –Yellow Pages (WaY) Sunday, August 16, 2015
Definitions Information retrieval (IR) the science of searching for information in documents, searching for documents themselves, searching for metadata which describe documents, or searching within (hypertext- )databases for text, sound, images or data. Knowledge Management (KM) refers to a range of practices used by organizations to identify, create, represent, and distribute knowledge for reuse and learning across the organization Gathering the information Doing something useful with it From:
The main Questions “Is semantic & concept search a competitive Edge?” “The role of the expert, trivial?” Sunday, August 16, 2015
The Case: TU Delft Library Goal: Online Information Literacy Instruction that students will actually use: Interactive Intuitive Fun What you see is what you get
Demonstration Sunday, August 16, 2015
Online user survey 12 respondents so far… Students (mostly BSc) and Professors 9/12 succeed a specific task using the Tag Cloud Search 8/12 like using it 8/12 think it’s useful
Future plans Can the Tag Cloud Search completely replace the traditional menu? Improve thesaurus based on user behaviour
Sunday, August 16, 2015