3.typing and advanced constructs
1 like495 views
The document discusses the use of semantic types, occurrences, and associations in topic maps for organizing unstructured data within the cloud platform Kamala. It explains how to create and manage topics, define relationships, and employ advanced constructs like scoping and reification to add structure to knowledge maps. The focus is on enhancing the semantic understanding of topics using URIs, merging related topics, and enriching knowledge representation.
3.typing and advanced constructs
- 1. Typing and advanced constructs Knowledge in the Cloud
- 2. Adding Seman3cs and Metadata to the unstructured data-‐ world 2
- 3. Typing ! Captures the seman3c knowledge in a topic map ▫ Topics (mul3ple types) ▫ Occurrences ▫ Associa3ons 3
- 4. Topic Types ! Reid Hoffman is a Person ! CEO is a Posi)on ! We model this as topic types of the corresponding topics CEO Person Posi3on Reid Hoffman 4
- 5. Topic Types ! Indicate that Reid Hoffman is an instance of the topic type Person ! This can be done at the topic details page ! When you first use Person as a type, Kamala asks whether to create a new topic ▫ ANer that you can reuse it 5
- 6. Topic Types ! Do the same for the topic on the Posi)on of CEO ! Define your own topic types and add them as type to the corresponding topics 6
- 7. Occurrence type ! The type of an occurrence indicates the seman3cs of the occurrence ! For example: ▫ Date of birth ▫ Email address ! Occurrence types are mandatory Reid Hoffman 1967-‐08-‐05 Date of birth 7
- 8. Associa4on Type ! Typing associa3ons is also mandatory ! Again, the type indicates the seman3c of the rela3on between the topics ! For example ▫ Employment L R Employment 8
- 9. Direc4on of Associa4on name Reid Hoffman LinkedIn Employed at Employed at?? Employs 9
- 10. Direc4on of Associa4on name Employs Employed at 10 ! Note: one of the names already exists ▫ We already defined it when crea3ng the first associa3on
- 11. Direc4on of Associa4on name ! Go to the employment topic by clicking on employs at the LinkedIn topic page ▫ Or search for employs ! Add the second name for the employment associa3on ! Add the full name: Employment ! Complete the names for all your associa3on types 11
- 12. Roles Walt Disney Company Pixar Subsidiary 12
- 13. Roles Pixar Organisa3on Subsidiary 13 Walt Disney Company Organisa3on
- 14. Roles Pixar Subsidiary Parent Daughter 14 Walt Disney Company
- 15. Roles W P Daughter Parent 15
- 16. Roles ! Think of appropriate roles for your associa3on types ! Note: in many cases role types are equal to the player types 16
- 17. Ontology page ! Now that our topic maps contain types, we can use them as entry-‐points ! You can use the ontology page for an overview at type level 17
- 18. Scoping, iden3fica3on, merging, reifica3on 18
- 19. Scope ! Scope adds seman3c context to topic characteris3cs ! For example: ▫ Mouse House is the nickname for Walt Disney ▫ BackRub was renamed to Google in 1998 ! Scope makes this difference possible ▫ Nickname can be a scope ▫ Past can be a scope ! Scoping is also oNen used for transla3ons 19
- 20. Scope ! Scope is a set of topics on ▫ Names ▫ Occurrences ▫ Associa3ons ▫ Roles ! Naming associa3on types is done by scoping ▫ The name for a direc3on is scoped by the appropriate role type topic 20
- 21. ! Experiment with scoping topic names in your topic map ! Add a scope to a name by hovering over the name and click Add scope ! The entry field that appears works just like any other topic field in Kamala Scope 21
- 22. Iden4fica4on ! A topic name is not necessarily unique ▫ The name Springfield is used for a lot of loca3ons and people hbp://en.wikipedia.org/wiki/Springfield Note that this URL is not a good iden3fica3on ! A name can have mul3ple meanings, such as spring ▫ The season ▫ Origin / source ▫ A verb: to rise 22
- 23. Iden4fica4on ! To uniquely iden3fy a topic, naming is not enough ! We use URIs ! Iden3fica3on makes it possible to recognize and merge topics describing the same subject 23
- 24. Iden4fica4on ! Go to hbp://www.linkedin.com/in/reidhoffman ! This page is about Reid Hoffman, execu3ve chairman of LinkedIn ! We can use this URI to iden3fy the topic on Reid Hoffman in our topic map 24
- 25. Iden4fica4on ! Add the URI as iden3fier via de topic page of Reid Hoffman 25
- 26. Iden4fica4on ! Do the same for the topic Execu)ve Chairman with the URL hbp://live.dbpedia.org/page/Chairman ! Now add other topic iden3fiers for your topics 26
- 27. Merging topics ! When two topics describe the same subject, they should be merged ▫ One topic equals one subject ▫ One subject equals one topic ! Make a new topic with the name Chairman ! This topic represents the same subject as the topic Execu)ve chairman and should therefore be merged ! By merging two topics, the proper3es of both topics are combined 27
- 28. Merging topics ! Go to the overview of Execu3ve Chairman ! Click the icon for merging ( ) ! Click Remember this topic on the tab By selec)on ! Go to the Chairman page and click the icon for merging ! Choose Merge remembered topic with this topic 28
- 29. Reifica4on ! What if we want to state something about an associa3on instead of a topic, for example the Employment of Reid Hoffman as execu3ve chairman ! Reifica3on is used to represent a topic name, associa3on or occurrence by a topic ! For the associa3on between Reid Hoffman and execu3ve chairman, you can state that this associa3on started in February 2007 29
- 30. Reifica4on Execu3ve Chairman Employment as Execu3ve Chairman Hoffman’s employment as EC February 2007 Start date 30 Reid Hoffman
- 31. Reifica4on ! Go to the topic Reid Hoffman ! At the role execu)ve chairman, hover over the arrow ! Choose Add reifica)on ! Name the new topic Hoffman’s employment as EC ! OK or enter and accept the crea3on of a new topic 31
- 32. Reifica4on ! Click on Reified as Hoffman’s employment as EC ! Add an occurrence of the type Start date and value February 1st 2007 with Add Quick Characteris3cs 32
- 33. Wrap up advanced constructs ! Using types, scopes, reifica3on and other constructs allows you to add structure and enrich your knowledge map ! View the next prac3cal session to learn more about ontology and schema
- 34. Thank you for your interest in Kamala. Keep up to date on new features at kamala-‐cloud.com 34