Understanding Semantic Mapping within STASIS

1
Understanding Semantic Mapping within STASIS
2
Example Min-max Inventory Replenishment
The supplier checks stocklevel periodically and
delivers his goods.
Supplier
Customer
Request Stock level
Deliveryadvise
Confirmation
3
Semantic
Informal specification based on terminology
Structural specification based on the data model,
i.e. XML Schema
Integrated method with sufficient formal
expressiveness
Ontologies
Common DatamodelOntology
BusinessOntologies
Folie 3
4
Format Example XSD
Folie 4
5
Common Data Model generic structure description

• Extensions for
• XML Schema
• Relational Databases
• EDI
• etc.

6
Describing the Schema Structure by using the CDM

• lt?xml version"1.0" encoding"UTF-8"?gt
• ltschema xmlns"http//..."gt
• ltelement namemyInvoice" type"invInvoiceTy
  pe"/gt
• ltcomplexType name"InvoiceType"gt
• ltattribute namemyDate"
  type"date"/gt
• ltattribute namemyCurrency"
  type"string"/gt
• lt/complexTypegt
• lt/schemagt

myInvoice
myCont_I
myDate
date
hasDataType
hasSourceNode
myCurrency
myCont_II
string
Folie 6
7
Schema Mapping
Folie 7
8
Describing Semantics by Annotation
Business Ontology
isA
isA
isA
isA
DateTimeUnit
FinancialUnit
Order
myInvoice
Invoice
myDate

myCurrency
yourInvoice

? Reasoning myInvoice equivalent to yourInvoice
Folie 8
9
Semantic Mapping
Folie 9
10
Automatic Mapping
Name
Complex Type
subclass_of
subclass_of
Buyer Name
Seller Name
is_a
is_a
myName
is_a
Possible automatic reasoning
yourBuyer
myFirstname
Mapping
myLastname
is_a
Concatenation
subclass_of
11
Combination
subclass_of
subclass_of
Name
Folie 10
11
Methods for automatic mapping 1/2

• String comparison
• good indicator for a perfect match
• Stem comparison
• improves the likelihood, but reduces confidence
• Levenshtein distance
• provide only a very week indicator for similarity
  
• Dictionary lookup
• gives a very high probability for a good match
• Abbreviation lookup
• may destroy valuable information

Folie 11
12
Methods for automatic mapping 2/2

• WordNet synsets
• used for further aggregation with scores from
  other plugins
• Structural comparison
• quality of this score heavily depends on the
  availability of metadata
• Transitive link reduction
• gives a strong indication, but only if all links
  in between are correct
• Reasoning
• provides the best indicator for a perfect match

Folie 12
13
Solution

• STASIS (SofTware for Ambient Semantic
  Interoperable Services)
• Aims to allow SME and corporate participation in
  the eEconomy
• Addresses semantic gap in integration
• Provides services and applications
• Industrial partners involved during requirements
  analysis, testing and validation

Folie 13
14
STASIS Core Components
Folie 14
15
STASIS Comparator - Parts for Annotation

• Provides a Wizard to load the user schema
• Provides a Wizard to load an Ontology
• Visualizes both
• User can annotate the elements in his Schema to
  an Ontology
• Again a Wizard is provided to help the user with
  annotation

Folie 15
16
Ontology Editor GUI
17
Mockup of the STASIS Annotation GUI
18
Project Information

• EU Funded project (6th Framework)
• 12 Partners across EU and China
• Commercial Universities User Organisations
• Links to standardisation and interoperability
  organisations
• CEN/eBES
• CEN/eBIF
• UN/CEFACT

Folie 18
19
Partner

• Commercial Companies
• TIE
• iSoft
• Universities
• Sunderland
• Oldenburg
• Modena e Reggio Emilia
• Tsinghua
• User Organisations
• AIDIMA
• Mariner
• Shanghai Sunline
• Foton
• TANET
• ZF Friedrichshafen AG

Folie 19