Why I Ignored Ontology and GraphDB and Why You Shouldn’t
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
Why I Ignored Ontology and
GraphDB and Why You Shouldn’t
Approach
•
•
•
Agenda
•
•
•
•
Excuse # 1
•
•
•
•*as of 2/15/2018
SPARQL
Excuse # 2
•
•
•
•
•
Excuse # 3
•
•
•
•
•
•
Excuse # 4
•
•
•
•
George Mary
George Dan
Mary Mark
Thomas George
hadChildWith
isParentOf
isParentOf
isParentOf
Family Person
(PK)ParentID(PK)PersonID
Name(PK, FK)ChildID
PersonIDName1092 George1492 Mary1826 Thomas3872 Mark4876 Dan
PersonParentID ChildID1826 10921092 38721092 4876
Family
Excuse # 5
•
•
•
•
Excuse # 6
•
•
•
•
•
•
Semantics and GraphDB Are:
•
•
•
•
Session Agenda
•
•
•
•
The Relational Approach I Grew up With
Conceptual
Logical
Physical
Relational – vs – Semantics
•
•
•
•
•
Human Resources
Point of Sale
Becky s Inventoryspreadsheet
Relational - vs Semantics
•
•
•
•
•
•
George Mary
George Dan
Mary Mark
Thomas George
hadChildWith
isParentOf
isParentOf
isParentOf
Relational - vs Semantics
••
•
•
••
•
•
Family Person
(PK)ParentID(PK)PersonID
Name(PK, FK)ChildID
PersonIDName1092 George1492 Mary1826 Thomas3872 Mark4876 Dan
PersonParentID ChildID
1826 10921092 38721092 48761492 38721492 4878
Family
George Mary
Dan Mark
hadChildWith
isParentOf isParentOf
Thomas
isParentOf
Relational - vs Semantics
•••••
•••
George Mary
George Dan
Mary Mark
Thomas George
hadChildWith
isParentOf
isParentOf
isParentOf
George Mary
Dan Mark
hadChildWith
isParentOf isParentOf
Thomas
isParentOf
Relational - vs Semantics
•••••
•••
George Mary
Dan Mark
hadChildWith
isParentOf isParentOf
Thomas
isParentOf
George Mary
Dan Mark
hadChildWith
isParentOf isParentOf
Thomas
isParentOf
isParentOfisParentOf
isSiblingOf
isGrandParentOf isGrandParentOf
Relational - vs Semantics
•
•
•
•
•
•
•
Relational - vs Semantics
•
•
•
•
•
Relational - vs Semantics
•
•
•
••
•
•
Relational - vs Semantics
•••••
•••
Subject ObjectPredicate
uriuri or
literal uri
ns:Subjectns:Object or literal
ns:Predicate
Point of Sale
Data
Accounting
Nobody Knows
Identity Mgmt
Becky s Inventoryspreadsheet
Payroll
Human Resources
Data Lake
Relational - vs Semantics
••
•
••
•
•
Point of Sale
Accounting
We Figured it Out
Identity Mgmt
Inventory
Payroll
Human Resources
Data lake
Relational - vs Semantics
••
•
••
•
•
The Semantics Approach = Simplicity
Conceptual
Logical
Physical
:Employee :Person::owl:subClassOf
:Employee :Person::owl:subClassOf
:Employee :Person::owl:subClassOf
Session Agenda
•
•
•
•
Before We Execute –What’s in a GraphDB?
••
•
•
••
•
•
Ontology & Graph DB: What is Different?
•
•
•
•
•
•
OWL Ontology Introduction
•
••
•
•
•
Kinds of Things vs. Individual Things
••
•
Rectangles for Kinds
(Ontology/Schema)
Rounded for Individuals
(Data)
No Thing is an Island
•
•
•
•
Inter-Related things in Healthcare
Dotted
lines can
be
inferred.
Healthcare in the Real World
Patient
Visit
SentaraOffer
Hospital
Contractor
Care
Provision
Outcome
Hospital
Accreditation
Hospital
License
Credentialing
Organization
giver
owns
Physical
Location
hasPhysical
Location
Sentara
OrganizationProcedure
Specification
Health Plan
Benefit regarding
Employee
Benefit
Building
occupies
Employee
hiresForWork
Employment
Contract
employs
Capability
hasA
Recommending
Treatment
Prescription
categorizedBy
governs
Payroll
Disbursement
Patient Bill
Membership
Premium
Health Product
Or Service
Delivery
Preference
Marketing
Message
Marketing
Communication
Sentara
Target Market
Business
Referral
refersTo
Insurance
Care
Resources &
Supplies
Employment
Underwriting
Claim
Insurance Product
regarding
Regulation
Regulator
produce
Sentara Healthcare Enterprise Ontology: Concise Overwiew
May 2013
Regulations
Supplier
Care Instructions
Marketing
Body of Knowledge
Legend
Class
SubClass
Class
Important SubClassClosely Related Class
Class
SubClass
Abstract Superclass
Another Classproperty
not explicit
Diagnosing
Medical Procedure
Medical
Treatment
Health Plan
Policy
Agreement
Health Insurance
Company
Patient
Registered
Patient
Person
Goal
BeHealthy
Image Goal
Provider Network
Diagnosis
Medical Condition
Disease
Physician
Healthcare
Provider
Nurse
Care Coordinator
Human Resource
Resource
Healthcare Facility
Equipment and
Tools
Supplier
Agreement
Sentara Vendor
Agreement
Pricing
Obligation
Volume
Obligation
In Hospital
Care Delivery
Mode
Physician Office
Patient Visit
Patient
Doctor
Inter-Related things in Finance
Things have a variety of relationships
•
•
•
Things have a variety of attributes
•
•
Example attributes
•
•
The Story So Far
•
•
•
•
Generalization and Specialization
•
•
Animal
Reptile
Primate Whale
Mammal
•
Finding “a kind of” relationships
•
•
••
•
•
•
Some “A Kind Of” Relationships
Drawing Conclusions – Traditionally
••
••
•
••
Drawing Conclusions – With Semantics
•
•
•
••
Two Kinds of Conclusions to Draw
NOTE: Don’t need
hard-coded rules.
Benefits of Drawing Conclusions
Logical Inconsistencies
•
Logical Inconsistencies
•
••
•
•
•
OWL as a Vocabulary for Data
•
••••••
OWL as a Vocabulary for Schema
•
••
•
•
•
•
•
Schema and Data Together in Graph DB!
••
•
•
••
•
••
•
NULL
Schema and Data Together in Graph DB!
••
•
•
••
•
••
•
NULL
Where do Triples Come From?
•
•
•
•
•
Thomas GeorgeisParentOf
How do we Build our Triples?*
•
•
•
•
•
•
Parent Child
Thomas George
George Mark
George Dan
Mary Mark
Mary Dan
Family
George MarkisParentOf
Session Agenda
•
•
•
•
Semantics Owns the Future, Will You?
••
•
••
•
••
•
Thank You!
:Presenter :Attendees:isGratefulTo
Related Documents
