Top Banner
Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1
33

Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

May 14, 2018

Download

Documents

votuyen
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Dr Ian PiperUK Director

PoolParty Semantic Suite

Content architecture, classification and knowledge graphs

1

Page 2: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Agenda

● Introducing SKOS taxonomies for classification● Modelling content and taxonomies for classification● Building content knowledge graphs

2

Page 3: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

(Part of) my knowledge graph

3

Page 4: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Introducing SKOSTaxonomies for classification

4

Page 5: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

SKOS is:

• The Simple Knowledge Organisation System

• A W3C standard: https://www.w3.org/TR/2009/REC-skos-reference-20090818/

• An RDF-conformant data model (an ontology)

5

(which may seem a little dry and dull)

Page 6: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

But really, and more importantly,SKOS is:

• A simple, well-defined, easy to understand format for classification schemes

• An extensible model for flexible taxonomy building

• A fundamental building block for semantic content classification

6

Page 7: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

So why is SKOS useful?

• It is easy to read• It provides basic taxonomy features

(broader, narrower, related, various labels)• Hierarchies are implied through broader-

narrower properties rather than explicit in element structures

• It is extensible (SKOS-XL, custom schemes)

7

Page 8: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Simplified SKOS model

• Two classes (types of thing): Concept and ConceptScheme

• These classes have properties• Some properties link concepts to other concepts• (via their URIs)• Other properties define and describe concepts

8

Page 9: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

9Another way to look at it

Page 10: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content and taxonomiesAligning content and taxonomies for classification

10

Page 11: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Old-school classification

• A traditional CMS taxonomy• Terms are simple strings – no extended

properties11

Page 12: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Old-school classification

• Traditional content tagged with taxonomy• The tag is a hard database relationship• No semantics (anywhere!)12

Page 13: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

What’s wrong with this?

• Not much, IF you have only simple content AND it’s all in one place AND your CMS and taxonomies never change

• It’s a fragile mechanism – it can break if you change anything

• It’s only a local solution – it won’t help you to connect up your enterprise

• Simple, monolithic content objects are inflexible – not easy to re-use

13

Page 14: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Another approach

• Think of content in terms of granular objects and containers

• Don’t worry about fixed storage hierarchies – use virtual hierarchies

• Give every content object a URI• Use external, enterprise taxonomies rather

than internal, local taxonomies• Build semantic links between content and

taxonomies

14

Page 15: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content objects and containers

15

Page 16: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content objects and containers

15

Page 17: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content objects and containers

15

Page 18: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content objects and containers

15

Page 19: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

A taxonomy in an external taxonomy management system

16Label and URI for this concept

Page 20: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

A piece of structured content with a URI

17The content object has a Uniform Resource Identifier (URI)

Page 21: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Putting it all together; a content object semantically tagged with a taxonomy concept

18

[has subject]

Page 22: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Simple content object structure

19

Page 23: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content model as a graph

20

Page 24: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content knowledge graphsHow taxonomies can work with structured content

21

Page 25: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Building a content knowledge graph - step 1

22

Page 26: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Building a content knowledge graph - step 2

23

Page 27: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Building a content knowledge graph - step 3

24

Page 28: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

25

An aside on semantic triples

Page 29: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Building a content knowledge graph - step 4

26

Page 30: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

27

Page 31: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

Content knowledge graphs: summary

A content knowledge graph approach:

• Allows separation of concerns and reduces dependencies

• Facilitates the development of an enterprise knowledge graph

• Provides an incremental route from current state

28

Page 32: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

CONNECT

Dr Ian PiperUK PoolParty DirectorTellura Information Services Ltd.• [email protected][email protected]• https://www.linkedin.com/in/ianpiper • https://twitter.com/tellura_tweets • http://tellura.co.uk/

29

© Semantic Web Company - http://www.semantic-web.at/ and http://www.poolparty.biz/

Page 33: Content knowledge graphs - informationenergy.org · Dr Ian Piper UK Director PoolParty Semantic Suite Content architecture, classification and knowledge graphs 1

CONNECT

Sebastian GablerConsultant, Semantic Web Company• [email protected]• https://at.linkedin.com/in/SebastianGabler• https://poolparty.biz

30

© Semantic Web Company - http://www.semantic-web.at/ and http://www.poolparty.biz/