2. Quick Agenda
2
! Problem: Why rating is not enough?
! Proposed Model
" Reputation Object (RO) Ontology
" Formal Model
! What technology to use?
" Output format example
! Related References
Reputation.Interoperability (Semantic Technologies);
3. Why Rating is not enough?
1. No context
3 E-Shop
Business Owner/
Relying Party Seller/Factory
User
4. Why Rating is not enough?
1. No context
3 E-Shop
Business Owner/
Relying Party Seller/Factory
Delivery
User Service
Delayed
Package
5. Why Rating is not enough?
1. No context
3 Bad E-Shop
Review
Business Owner/
Relying Party Seller/Factory
Delivery
User Service
Delayed
Package
6. Why Rating is not enough?
1. No context
3 Bad E-Shop
Review
Business Owner/
Relying Party Seller/Factory
Delivery
User Service
Delayed
Package
Context excluded from the reputation value
" reputation query is too general
" key components contexts (i.e. quality aspects)
" delivery, quality, price
7. Why Rating is not enough?
2. Different perceptions
4
Different representations, interaction styles and
trust rating scales
8. Why Rating is not enough?
2. Different perceptions
4
Isolated reputation communities that have different:
" perception of reputation
" calculation of reputation
" interpretation of reputation
" overall reputation – not context related
Different representations, interaction styles and
trust rating scales
9. Why Rating is not enough?
3. No portability
5
" Starting from scratch for each domain
" Cold start problem
" No reputation information exchange
10. Why Rating is not enough?
3. No portability
5
" Starting from scratch for each domain
" Cold start problem
" No reputation information exchange
Solution
!Unify the representation not the calculation
!Facilitate knowledge exchange
11. Reputation Object Model
Representation
6
Reputation
is the notion of profiling an entity’s
performance
12. Reputation Object Model
Representation
6 The proposed RO model
! Uses more information about the domain
" the contexts and relevant quality criteria
! Using this information, reputation is represented differently
" as a developed object
! The Reputation Object profiles an entity’s performance and has knowledge about
" contexts
" ratings values/reviews/feedback
" computation functions
" collecting method
Reputation
is the notion of profiling an entity’s
performance
17. selection, usage control, cloud service selection, and in a rule-based open reputation
system. I also give a summary on formalizing the model and implementation. Four joint
publications are submitted and listed in the references section.
The Formal Model
1 RO Model Formalism and Implementation
8
Definition ROs represents all reputation objects and is defined by:
ROs = (A, C, R, range, rep, order), where
• A is the set of all the entities that can have a reputation or can be evaluated.
• C is the set of criteria/context (relevant category in which a reputation is earned).
• range maps the criteria to its possible values, such that: Let P (V ) be the set of all
possible values that a criterion/context c ∈ C can have; then range : C → P (V )
• R is the set that contains the relative pairs of (entity, criteria) only, such that:
R⊆A×C
• rep maps a criteria to its value after a new rating transaction, where rep : R → V
such that: rep(a, c) ∈ range(c), a ∈ A
• order maps the set of possible values P (V ) to its relevant order and is used in the
comparison between two given values in the P (V ) set, where order : C → P (V 2 )
1.1 Ontology and Implementation
Reputation.Interoperability (Semantic Technologies);
18. Which Technology?
Used Technology
9
! Developing interoperable reputation objects requires a
technology that can:
! structure and standardize reputation info and its relevant data
! enable data integration
! provide ways to relate the data to its semantics
Reputation.Interoperability (Semantic Technologies);
19. Which Technology?
Used Technology
9
! Developing interoperable reputation objects requires a
technology that can:
! structure and standardize reputation info and its relevant data
! enable data integration
! provide ways to relate the data to its semantics
Semantic
Technologies
Reputation.Interoperability (Semantic Technologies);
20. Option 1: RDF graphs
RO as RDF graph
10 • “Bob has a very good Delivery”
• Service Quality=0.87
• Payment.method
Reputation.Interoperability (Semantic Technologies);
21. Option 1: RDF graphs
RO as RDF graph
10 • “Bob has a very good Delivery”
• Service Quality=0.87
• Payment.method
Table 1: Reputation Statements about Bob
Target Criterion Value
Bob Service Quality 0.87
Bob Delivery ”very good”
Bob Payment <purl.org/goodrelations/v1/MasterCard>
8B(C8
of statements correspond to the RDF statements (or triple)
!""#$%%+=&(*'%>?-,.":
form of: <subject,predicate,object>, where the reputa-
5&67
<(-<$/-A+ tion statement in this case is: <target,context,value>.
The same as an RDF graph which is a set of RDF triples,
!""#$%%+=&(*'%)+,.1+*: reputation statements therefore form a reputation
the set of
RDF graph.89+*:;<-0"8
Lets assume that we are rating a seller in an e-
market identified by <foaf:Person rdf:nodeID="Bob"> then
!""#$%%+=&(*'%@-:A+/"
a simple description of his reputation can be viewed as declar-
ing the statements in table 1. If Bob’s servie-quality, deliv-
!""#$%%&&&&'(()*+,-".(/0%
12%3-0"+*4-*) ery, and payment are identified by URIs as well as the literal
Reputation.Interoperability (Semantic Technologies);and ”very good”, this table corresponds to the
values 0.87
22. Option 2: Ontology for Expressiveness
RO Ontology
11 ! Developed using Protégé 3.4.4 OWL
! Integration on the implementation layer: Java library
! Vocabulary of RO Ontology:
! to represent an entity's (foaf:Agent) reputation
! an object (ReputationObject) has one or multiple
instances of class Criterion or QualityAttribute
! each criterion instance has a ReputationValue
(currentValue and historyList) that has a set of
PossibleValues (as literals or resources URI)
! a criterion is collected by a CollectingAlgorithm &
computed using a ComputationAlgorithm
! Employing also known vocabulary
OWL, RDFS, FOAF, XSD, RDF Review, ..
23. Using Semantic Technologies
Goals
12
! enabling reputation information exchange
! facilitate the integration of multiple sources to draw new conclusions,
! connecting data to its definitions and to its context
! achieving reputation interoperability
! Context-aware reputation
! ensuring understandability and reusability of the embedded information
Semantic
Technologies
24. PossibleValues, describes the order of the possible values for a criterion OWLList
to be able to compare between 2 values
PossibleValues, describes the comparison function (i.e. between two Algorithm
Output Format in XML/OWL
given reputation values) and is used as an alternative to order a dy-
namic set of possible values if a static list is not given
Criterion or QualityAttribute Applications
ComputationAlgorithm
13 Example: A seller
Criterion or QualityAttribute RO in e-Markets CollectingAlgorithm
Rating type:literal
! Using GoodRelations ontologies to describe a seller and RO ontology to
describe its reputation
eResponder rule-based agents can Listing 3: Seller’s RO
other and can exchange reputation <gr:Reseller rdf:reference=”http://www.example.org/John#” >
t-output values of the reputations <ro:hasReputation >
<ro:ReputationObject rdf:ID=’’SellerRO1’’>
see listing 2). [25]
<ro:hasCriteria>
mentation of decentralized reputa- <ro:Criterion rdf:resource=’’http://purl.org/goodrelations/v1/
ts interchange and evaluate their DeliveryMethod’’>
lso for centralized reputation mod-
Criterion 1 <ro:hasReputationValue>standard</ro:hasReputationValue>
service nodes act as trusted repu- <ro:collectedBy ro:CollectingAlgorithm=’’#WebPortal’’/>
em. The interchanged reputation </ro:Criterion>
<ro:Criterion>
the internal rule-based decisioning
<review:Review>
policies of a RuleResponder agent. <review:rating>8</review:rating>
Criterion 2
ight give certain rights to a trusted </review:Review>
</ro:Criterion>
</ro:ReputationObject>
</ro:hasReputation >
unicate Reputation Objects </gr:Reseller>
nt,acl query−ref, QueryRO), The decision rule of a customers’ agent to buy a product
,acl inform−ref, ReceivedRO), (e.g. a book) from a certain seller depends on the rating of
25. …finally
14 Solution
!Unify the representation not the calculation
!Facilitate knowledge exchange
Reputation.Interoperability (Semantic Technologies);
26. …finally
14 Solution
!Unify the representation not the calculation
Reputation
!Facilitate knowledge exchange
is the notion of profiling an entity’s performance
Reputation.Interoperability (Semantic Technologies);
27. …finally
14 Solution
!Unify the representation not the calculation
Reputation
!Facilitate ! Developing interoperable reputation
knowledge exchange
is the notion of profiling an entity’s performance
objects requires a technology that can:
! structure and standardize reputation
info and its relevant data
! enable data integration
! provide ways to relate the data to its
semantics
Reputation.Interoperability (Semantic Technologies);
28. …finally
14 Solution
!Unify the representation not the calculation
Reputation
!Facilitate ! Developing interoperable reputation
knowledge exchange
is the notion of profiling an entity’s performance
objects requires a technology that can:
! structure and standardize reputation
info and its relevant data
! enable data integration
! provide ways to relate the data to its
Solution
semantics !Semantic Technologies
Reputation.Interoperability (Semantic Technologies);
29. …finally
14 Solution
!Unify the representation not the calculation
Reputation
!Facilitate ! Developing interoperable reputation
knowledge exchange
is the notion of profiling an entity’s performance
objects requires a technology that can:
! structure and standardize reputation
info and its relevant data
! enable data integration
! provide ways to relate the data to its
Solution
semantics !Semantic Technologies
Reputation Object as a semantic knowledge
object (not just simple rating)
Reputation.Interoperability (Semantic Technologies);
30. …finally
14 Solution
!Unify the representation not the calculation
Reputation
!Facilitate ! Developing interoperable reputation
knowledge exchange
is the notion of profiling an entity’s performance
objects requires a technology that can:
The model therefore achieves:
! structure and standardize reputation
! the reputation of an entity is more meaningful
info and its relevant data
! associated with the context in which it was earned
! enable data integration
! automation of criteria assignment is possible by declaring a relevant
resource as aprovide ways to relate the data to its
! criterion
Solution
! (ex. URI1 is_a _:criterion)
semantics !Semantic Technologies
Reputation Object as a semantic knowledge
! one can easily extend these object (not just simple rating)the list of
criteria dynamically by adding to
contexts/criteria in the reputation objects
Reputation.Interoperability (Semantic Technologies);
31. …finally
14 Solution
!Unify the representation not the calculation
Reputation
!Facilitate ! Developing interoperable reputation
knowledge exchange
is the notion of profiling an entity’s performance
objects requires a technology that can:
The model therefore achieves:
! structure and standardize reputation
! the reputation of an entity is more meaningful
info and its relevant data
! associated with the context in which it was earned
! enable data integration
! automation of criteria assignment is possible by declaring a relevant
resource as aprovide ways to relate the data to its
! criterion
Solution
! (ex. URI1 is_a _:criterion)
semantics !Semantic Technologies
Reputation Object as a semantic knowledge
! one can easily extend these object (not just simple rating)the list of
criteria dynamically by adding to
contexts/criteria in the reputation objects
Usability? The degree of visibility
for these criteria depends on the
web site
Reputation.Interoperability (Semantic Technologies);
32. Related References
15
! Rehab Alnemr, Adrian Paschke, Christoph Meinel, "Enabling Reputation
Interoperability through Semantic Technologies", ACM International
Conference on Semantic Systems, Sept 2010.
! Rehab Alnemr, Stefan Koenig, T. Eymann and C. Meinel, "Enabling Usage
control through Reputation Objects: A discussion on e-Commerce and the
Internet of Services environments", in the special issue of Trust and Trust
Management, Journal of Theoretical and Applied Electronic Commerce
Research, 2010.
! Rehab Alnemr, Christoph Meinel,"From Reputation Models and Systems to
Reputation Ontologies", Proc. 5th IFIPTM, Springer IFIP, Copenhagen,
Denmark, July 2011 (to appear)
! Adrian Paschke, Rehab Alnemr, Christoph Meinel, "The Rule Responder
Distributed reputation Management System for the Semantic Web",
RuleML-2010 Challenge, Washington DC, USA.
Reputation.Interoperability (Semantic Technologies);