Deals about how the web services working on semantic matching

1. SEMANTIC MATCHING OF
WEB SERVICES
CAPABILITIES
- Velkumaran A P (17PW38)

2. TABLE OF CONTENTS
01
Semantic
Matching
03
DAML-S
Profiles
02
Matching
Engine
04
Matching
Algorithm

3. Web Service provide an new model of
the web in which sites exchange
dynamic information on demand
Locate and Interoperate
Semantic Matching

4. DAML - S : Capabilities, Specification of Matching algo
It provides semantically based view of web services
which spans from the abstract description of the
capabilities of the service to the specification of the
service interaction protocol, to the actual message that
it exchanges with other web services.

5. Why SOAP, WSDL, UDDI can’t use for location?
SOAP and WSDL - Description of the message, transport
mechanisms and describing the interface
UDDI - describes businesses by their physical attributes(name,
address and services). TModels(Classification of the services
within taxonomies). But it doesn’t provide an services
capabilities, no use.
Two identical XML descriptions may mean very different things
depending on the context of their use

6. DAML + OIL
Semantic representation of services
Subsumption reasoning on taxonomies
Allows the definition of relations between concepts
Lack of definition of well formed formulae and an
associated theorem

7. DAML-S Profiles
Functional
Attributes
Functional
Description
Inputs, Outputs,
Precondition
Actor
Stores the information of
the provider
Quality ratings

8. Describes the functionalities that a web service wants to
provide to the community.
Web services have many functionalities.
Example: Book selling service
Browse and Buy -> Functionality to advertise determines how
the service will be used

10. MATCHING ENGINE
We envision a Web Wide Infrastructure for web services supported by a set of registries that functions
as directories.
Degree of similarities
Sufficiently similar -> flexible matches
Service requesters should also be allowed to decide the degree of flexibility, they reduce the likelihood
of finding services that match their requirements.
Exploitation - Too generic in the attempt to maximize the likelihood of matching

12. Semantic matches despite of syntax difference and modelling abstraction
DAML also supports accuracy
No matching is recognized when the relation between the advertisement and the
requester does not derive from the DAML ontologies.
The semantic of DAML-S descriptions allows us to define the ranking function.

13. MATCHING ALGORITHM
Request is matched against all the advertisement and stored it.
Rules for matching each advertisement with the request
Degree of Matching
Sorting the matches

14. MAIN CONTROL LOOP

15. OUTPUT MATCHING

16. DEGREE OF MATCHING ASSIGNMENT
Degree of Matching is determined by the
minimal distance between the concepts in
the taxonomy tree.
Four degree of Matching

17. Exact
If OutR = OutA (or) If OutR subclass of OutA
PlugIn
If OutA subsumes OutR
The rule acknowledges there is weaker relation between OutR and OutA
We can except that a service that advertises an output of vehicle provides some types of cars,
but we cannot except that it provides ever types of SUV
Subsumes
If OutR subsumes OutA, then the provider cannot fulfill the request.
Failure
Failure occurs when no subsumption relation between advertisement and request is identified.

18. SCORING SYSTEM

19. PROBLEM : LOOKING FOR CARS
The service is advertised is a car selling service which given a price reports which car
can be bought for that price

20. Exact Match because it subsumes but it will be a Plugin if the Advertisement output is
Vehicle instead of Car

21. The only thing that matters during matching is whether the matching engine can draw a
inference between inputs and outputs of the advertisement and request based on the
ontology.
Not boolean values but the degree of similarity
Minimal acceptance degree
Constraint of amount of search

22. THANKS FOR WATCHING !!