Weitere ähnliche Inhalte Ähnlich wie Semantische Technologien (nicht nur) für die verbesserte Suche in SharePoint (20) Kürzlich hochgeladen (20) Semantische Technologien (nicht nur) für die verbesserte Suche in SharePoint1. Semantische Technologien
(nicht nur) für die verbesserte
Suche in SharePoint
Daniel Hansch
Shared Solutions Day – 20. Februar 2014
DIQA Projektmanagement GmbH
Pfinztalstraße 90
76227 Karlsruhe
info@diqa-pm.com
2. About DIQA GmbH
DIQA is an independent software vendor of knowledge management tools for
ECM portals.
Our vision:
We provide our customers with services and products that turn their ECM
portals into smart portals by introducing semantic web technologies. Smart
portals let end-users better find, organize, process, control and govern
unstructured content.
Founded:
Team:
Location:
DIQA Portfolio, January 2013
2012
SharePoint, MediaWiki, knowledge management and semantic
web specialists
Germany, Karlsruhe
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 2
3. Agenda
• The Semantic Web
•
•
•
•
Vision, Goals
Principles
Base technologies
Available data
•
•
•
•
BBC Semantic Publishing
Google Knowledge Graph
Facebook Open Graph
Wikidata
• Applications:
• Using the Semantic Web in SharePoint
• Semantic Search in SharePoint
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 3
4. The Semantic Web
• Tim Berners-Lee’s vision of a semantic web:
The Semantic Web isn't just about putting data on
the web. It is about making links, so that a person or
machine can explore the web of data. With linked
data, when you have some of it, you can find
other, related, data.
http://www.w3.org/DesignIssues/LinkedData.html
• Note: We treat the terms as synonym:
• Semantic Web
• Web of Data
• Linked (Open) Data
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 4
5. Linked Data Principles
★
★★
★★★
★★★★
Available on the web (whatever format)
… with an open license, to be Open Data
Available as machine-readable structured data (e.g.
excel instead of image scan of a table)
Available in a non-proprietary format (e.g. CSV
instead of excel)
Using open standards from W3C (RDF and SPARQL) to
identify things, so that people can point at your stuff
★★★★★ Linked to other people’s data to provide context
Tim Berners Lee (2010): http://www.w3.org/DesignIssues/LinkedData.html
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 5
6. RDF Data Model
• Web of Data is based on
RDF data model
• RDF is a semi-structure
graph data model
• Nodes and edges are
labeled with URIs
• Basic pattern (triple)
• subject-predicate-object
• BusinessEntity1 offers Offering1
• UnitPriceSpec1 hasValue “200.0”
• RDF can be serialized in
many formats, incl.
RDF/XML
http://www.heppnetz.de/projects/goodrelations/primer/images/fig1.png
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 6
7. Linked Data Cloud 2007
Source for this and the folllowing graphs: Linking Open Data cloud: Richard Cyganiak, Anja Jentzsch
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 7
8. Linked Data Cloud 2008
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 8
9. Linked Data Cloud 2009
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 9
10. Linked Data Cloud 2010
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 10
11. Linked Data Cloud 2011
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 11
12. Agenda
• The Semantic Web
•
•
•
•
Vision, Goals
Principles
Base technologies
Available data
•
•
•
•
BBC Semantic Publishing
Google Knowledge Graph
Facebook Open Graph
Wikidata
• Applications
• Using the Semantic Web in SharePoint
• Semantic Search in SharePoint
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 12
13. Linked Data Cloud 2011
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 13
14. BBC
Early adopter of the WoD („Linking open data project“), roles:
• Data provider (program catalogue, artists)
• Data consumer (links to external resources about artists)
• Technology provider (similar to Thomson Reuters, Elsevier and NYT?)
Dynamic Semantic Publishing architecture
• Semantic web technology stack to reduce curation effort for online media
production
• Challenge: BBC Sports sites for 2010 World cup, Olympic games: 700 index
pages require curation, like links to story pages etc. and frequent updates.
• DSP replaces static publishing with dynamic aggregation that makes use of a
metadata layer.
• Workflow:
• Editors author stories
• Stories are tagged (semi-)automatically
• Index pages are generated automatically and kept up-to-date through
queries that use tags.
Benefit
• Reduced effort for curation
• Deeper and broader access to BBC content
• Increased quality
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 14
19. Agenda
• The Semantic Web
•
•
•
•
Vision, Goals
Principles
Base technologies
Available data
•
•
•
•
BBC Semantic Publishing
Google Knowledge Graph
Facebook Open Graph
Wikidata
• Applications
• Using the Semantic Web in SharePoint
• Semantic Search in SharePoint
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 19
20. Google Knowledge Graph
• 2005 Google hires Guha (co-inventor of RSS and RDF)
• 2010 Google acquires Metaweb (developers of
Freebase)
• 2011 Bing, Google and Yahoo! introduced
Schema.org.
• Goal: common set of schemas for structured data
markup on web pages
• Based on ontologies and formal metadata
• Improve Search results
• 2012 Google starts enhancing search results with
formal metadata from the Knowledge Graph
• Based on wikipedia-crawls (~DBPedia)
• Freebase
• CIA World Factbook and more
• 2013 Google hires Denny Vrandecic (co-inventor of
Semantic MediaWiki and Wikidata) …
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 20
22. Facebook Open Graph
• Started as the Social Graph (friends)
• Now, every web-page/thing can become a node in the
Facebook Graph
• Social plugins on pages, e.g. Like
• Nodes can be linked with different kinds of edges
• Friend, Like, write, listen, eat, cook
• Graph API makes data readable and writable for Facebook
Apps
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 22
25. Agenda
• The Semantic Web
•
•
•
•
Vision, Goals
Principles
Base technologies
Available data
•
•
•
•
BBC Semantic Publishing
Google Knowledge Graph
Facebook Open Graph
Wikidata
• Applications
• Using the Semantic Web in SharePoint
• Semantic Search in SharePoint
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 25
26. Linked Data Cloud: Life Sciences Data
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 26
27. Other Sources for data in Life Sciences
• From the LOD cloud
• UniProt
• SIDER
• DrugBank
• PubMed
• GeneOntology
• PubChem
• ChEMBL
• KEGG Drug, Pathway,
Enzyme, Reaction, …
• …
• LinkedLifeData combines
• ChemBI
• DiseaseSome
• DrugBank
• EntrezGene
• GeneOntology
• NCI
• SIDER
• PubMed
• UMLS
• Uniprot
• …
http://linkedlifedata.com/
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 27
28. Use Linked Data from Uniprot to Filter SharePoint
Documents
Terms from Uniprot are used as
“Semantic Tags”. Each tags is associated
with an enzyme in Uniprot. This list of
documents is generated from a SPARQLquery that returns all documents about
an enzyme, that has “Magnesium” as
cofactor.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 28
29. SharePoint add-on from DIQA: GRASP
GRASP accesses SPARQL
endpoints from the web of data.
GRASP Visualizations in
Web Browser
1)
GRASP
SPARQL
SharePoint 2010
Read more about GRASP:
http://www.diqa-pm.com/en/GRASP
1) Linking Open Data cloud: Richard Cyganiak, Anja Jentzsch
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 29
30. Agenda
• The Semantic Web
•
•
•
•
Vision, Goals
Principles
Base technologies
Available data
•
•
•
•
BBC Semantic Publishing
Google Knowledge Graph
Facebook Open Graph
Wikidata
• Applications
• Using the Semantic Web in SharePoint
• Semantic Search in SharePoint: SharePoint
Findability Solution
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 30
31. DIQA‘ S S HARE P OINT F INDABILITY S OLUTION
• TERMINOLOGY MANAGEMENT
• AUTOMATIC DOCUMENT CLASSIFICATION
• INTELLIGENT SEARCH
DIQA Projektmanagement GmbH
Pfinztalstraße 90
76227 Karlsruhe
info@diqa-pm.com
32. SharePoint Findability Solution: Features
1.
2.
3.
4.
5.
6.
7.
8.
Upload and manage terminologies in the “library of
ontologies” (e.g. SKOS and TBX/TermBase eXchange).
Load terminologies into term stores, groups or term sets.
Manage the terms in the terminology manager (e.g.
labels in different languages).
Manage the relations between terms including
associations and poly-hierarchies.
Create classification rules in order to automatically tag the
document corpus (requires Layer2 Autotagger).
Use the terminology to intelligently suggest search terms in
the document search (Term Suggester).
Use the TreeView Refiner to drill-down or drill-up in the
search results.
The user is guided in the search process by the „Matching
Terms“ and „Related Terms“ webparts.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 32
33. 1. Library of ontologies
http://server/
Upload
terminologies (in
SKOS or TBX) and
manage them in
a library.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 33
34. 2. Load terminologies into the termstore
http://server/
1. Select a
terminology or
taxonomy to
populate a term
store…
2. Select the term
store and the
update strategy.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 34
35. 3. Manage terms
DIQA Portfolio, January 2013
Manage term
labels in different
languages,
descriptions, …
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 35
36. 4. Manage relations between terms
Add terms that
are related to this
term…
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 36
37. 4. Manage relations between terms
Manage multiple
parent terms (poly
hierarchy)…
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 37
38. 4. Manage relations between terms
…pick parent
terms from the
tree browser.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 38
39. 4. Manage relations between terms
Inspect the full
term hierarchy in
the TreeBrowser.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 39
40. 5. Define classification rules
If a document
satisfies this rule
then it is tagged
with a specific
term.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 40
41. 5. Define classification rules
Validate the rule
before it is used to
analyze your
entire document
corpus.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 41
42. 5. Tag documents automatically
Entire SharePoint
content is tagged
automatically
based on the
classification rules.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 42
43. 6. Search terms are intelligently suggested
The Term Suggester
Webpart supports
the user while he is
typing in his search
query…
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 43
44. 6. Search terms are intelligently suggested
…the intelligent
matching algorithm
suggests terms from
the terminology that
contain parts of the
search query in
labels and
synonyms.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 44
45. 7. Term-tree to navigate in search results
TreeView Refiner
Webpart extends
the standard refiner
webpart and
visualises the terms
in the context of the
term-tree.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 45
46. 7. Term-tree to navigate in search results
Users can select
terms in the termtree to drill down or
drill up in the search
results.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 46
47. 7. Term-tree to navigate in search results
DIQA Portfolio, January 2013
Search results
are updated
as you
navigate in the
© 2013 DIQA Projektmanagement term tree. | Slide 47
GmbH | www.diqa-pm.com
48. 8. Matching terms guide the user in the search process
Pick a new search
term from the list of
matching terms
and resume the
search.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 48
49. Advantage over standard SharePoint-Search
1. Superior managed metadata for content classification
2. Integrated taxonomies from various sources
3. Reliable automatic document-tagging
4. Users find documents immediately despite unknown
taxonomy
5. Users are guided in the search process
6. The terms contained in the search results are presented in
their taxonomic context
7. Users can easily drill-up or drill-down in the tree to broaden
or narrow the search
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 49
51. Take Home Message
• Semantic Web
• Open standards for publishing structure data
(graph knowledge)
• Vast number of available data sources
• DIQA makes this knowledge accessible in
SharePoint
• Metadata is one key benefit of SharePoint
Stop searching, start finding: the "SharePoint
Findability" solution from DIQA provides reliable
products and a proven method to find
documents quicker and more efficiently.
DIQA Portfolio, January 2013
© 2013 DIQA Projektmanagement GmbH | www.diqa-pm.com | Slide 51
52. Thank you for your attention!
Visit us on http://www.diqa-pm.com
DIQA Projektmanagement GmbH
Pfinztalstraße 90
76227 Karlsruhe
info@diqa-pm.com