Towards an Ontology for Ubiquitous User Modeling Interoperability
María de Lourdes Martínez-Villaseñor
1,2
and Miguel González-Mendoza
2
1
Universidad Panamericana, Campus México, Augusto Rodin 498, Col. Insurgentes Mixcoac, Mexico D.F., Mexico
2
Instituto Tecnológico y de Estudios Superiores de Monterrey, Carretera Lago de Guadalupe Km 2.5,
Atizapán de Zaragoza, Edo. de Mexico, Mexico
Keywords: Ontology, User Modeling, Interoperability, Ubiquitous User Model, Sharing and Reusing Information,
SKOS.
Abstract: In order to obtain a broader understanding of the user, some researchers in the community of user modeling
envision the need to share information of user models between applications. But gathering distributed user
information from heterogeneous sources to obtain user models interoperability implies handling syntactic
and semantic heterogeneity. It is also important to provide means for a ubiquitous user model to evolve over
time. We present U2MIO a dynamic ontology with flexible structure for user modeling interoperability
based in SKOS ontology. The U2MIO provides mediation based user modeling for sharing and reusing
information from heterogeneous user models. A two-tier matching strategy is proposed for the process of
concept alignment that permits the interoperability between profile suppliers and consumers.
1 INTRODUCTION
Today, users actively interact with systems and
applications through several devices. Many service
providers develop user-adaptive systems and are
trying to model the user for adaptation and
customization. But these efforts are frequently
isolated and repeated by each supplier; each of them
obtains a narrow understanding of the user. Some
researchers in the community of user modeling
(Heckmann, 2005) (Houben et al., 2005) Berkovsky
et al., 2007) (Carmagnola, 2009) envision the need
to share information of user models between
applications to achieve a broader understanding of
the user in order to provide better service, content
and interface personalization and adaptation.
Carmagnola (Carmagnola, 2009) denote “the
process of gathering and making sense of distributed
user information as user models interoperability and
the systems exchanging knowledge about a user as
interoperable user-adaptive systems”. Sharing and
reusing information between models can bring the
following advantages:
Prevent the user to repeatedly setting
parameters in devices, applications and
services which provokes inconsistency,
redundancy and repeated configurations.
Helps deal with “cold start problem”
(Leonardi, 2009) when a user is new to
applications and services.
Provides enrichment to the user models
obtaining a better understanding of the user
having a broader picture of user’s
characteristics, preferences and interest as well
as user’s current state.
But gathering distributed user information from
heterogeneous sources to obtain user models
interoperability implies handling syntactic and
semantic heterogeneity. Several techniques have
been proposed to deal with the lack of
interoperability (Viviani et al., 2010). The two major
approaches for user modeling interoperability are the
definition of standard ontologies that provide a
common understanding among multiple systems or a
conversion approach using mediation techniques to
provide mappings between the different user models.
Both approaches try to deal with syntactic and
semantic heterogeneity in different manners, but it
has become clear in recent years, that the user model
representation must be able to evolve over time,
must avoid redundancy, and provide security
management. In this paper, we address the syntactic
and semantic heterogeneity between the sources, and
the evolutivity problems of the user model.
We present a dynamic ontology with flexible
structure based in Simple Knowledge Organization
239
de Lourdes Martínez-Villaseñor M. and González-Mendoza M..
Towards an Ontology for Ubiquitous User Modeling Interoperability.
DOI: 10.5220/0004141202390244
In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2012), pages 239-244
ISBN: 978-989-8565-30-3
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
System (SKOS) (Miles et al., 2007), to provide
knowledge representation for the construction of a
ubiquitous user model in order to enable user
modeling interoperability. A process of concept
alignment determines the semantic mappings
between the suppliers and consumers’ concept
scheme and the ontology and gives
recommendations to change its structure.
Related work is discussed in section 2. The
ubiquitous user model interoperability ontology
(U2MIO) development is explained describing the
processes of ontology set up, ontology manipulation
and concept alignment for interoperability in
heterogeneous user models (section 3) . We compare
our approach with related work in 4 and conclude in
the final section.
2 RELATED WORK
One frequently cited representative of
standardization-based user modeling ontologies is
GUMO, developed for ubiquitous user modeling. In
(Heckman, 2005) the author describes general user
model ontology (GUMO) and its basic user
dimensions. This ontology is intended for the
description of the user major dimensions and
situational statements with the user model
auxiliaries, predicate and ranges. However GUMO
user model is almost static, although well suited to
represent situational statements.
Some recent approaches (Sutterer, 2009) are
based on user profile management selection taking
into account user profile management standards and
specifications like (ETSI, 2005) (3GPP GUP, 2009).
Sutterer presented a user profile management
framework for service platforms. The profile
structure that includes situation-specific user sub-
profiles for each application considered is interesting
because it adapts to changing situations and is
flexible to include new applications. It aggregates
profiles and sub-profiles for each application and
selects the best suited sub-profile for each situation
related to a service platform. It also permits the
aggregation of user models, these causes redundancy
when saving the same profile structure with minor
changes several times.
Iqbal et al. (Iqbal et al., 2010) (Iqbal et al., 2011)
presents a privacy-enhanced innovation framework
designed considering standard semantic web tools
for interoperability and integration of data from
multiple sources like social networks,Friend of a
Friend (FOAF) (Brickley, 2007). Their approach is
important in the matter of dealing with personal and
corporate issues to foster enterprise collaboration.
They also set up the personal profile extracting
information from heterogeneous sources, but they do
not provide mapping between the sources in order to
permit automatic reuse and sharing of individual
concepts.
The Generic User model Component (GUC)
presented in (van der Sluijs and Houben, 2006)
proposed to construct combined ontologies for
exchanging user models between web-based
systems. GUC allows the configuration of a
distributed management of mappings between user
models. The schema mappings were determined by a
human and the possibility of automatic merging
techniques was discussed but implemented requiring
human effort. Carmagnola et al., (Carmagnola et al.,
2009) proposed a solution with high flexibility
representing user models and providing semantic
mapping of the user data from heterogeneous
sources. However, to take part in the interoperability
process every provider needs to comply with a
standard format for the exchange and maintain a
sharable user model.
3 U2MIO ONTOLOGY
In this section, we first outline relevant aspects for
the ubiquitous user model interoperability ontology
(U2MIO) development. Subsequently, we describe
the processes of ontology set up, manipulation and
the concept alignment for interoperability in
heterogeneous user models.
3.1 Purpose and Scope of the Ontology
The ubiquitous user model is part of a user-adaptive
system that will facilitate knowledge sharing and
reuse between heterogeneous user profiles. The aim
of ubiquitous user modeling interoperability
ontology is to provide a shared and common
understanding between profile suppliers and
consumers in order to achieve cross-system
personalization and sharing and reuse of user
models. We are considering social network profiles,
data from advanced sensors attached to personal
gadgets, semantic web technologies like FOAF and
personal health records as profile suppliers. Some of
these devices and applications can be profile
consumers also. Other profile consumers considered
are web services, as we intend to use the information
as parameters of web service in order to help
personalization. The main problems that the
ontology is presumably solving are:
KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
240
To support semantic interoperability
overcoming differences between concepts at
knowledge level.
Represent a flexible user profile structure, with
domain independency which provides the
possibility for the ubiquitous user model to
evolve during time.
It must support that new profile suppliers and
consumers take part in the interoperability
process without effort of the provider or
consumer system.
The adaptive system must be able to contemplate
new sources, applications, devices and web services.
Given the semantic diversity of the heterogeneous
user models considered (and to come) as profile
suppliers and consumers, a flexible ontology that
permits “.. ongoing modeling, ongoing sharing and
ongoing exploitation” (Heckmann, 2005) is needed.
The system’s end user is any person using social
network applications, personal gadgets with sensors,
personal health records, FOAF and web services that
would like to avoid repeatedly setting his profiles,
by letting the system automatically set and enhance
them as well as automatically use this information
for adaptation and personalization, in particular web
service personalization. The ubiquitous user model
information could be useful in content and interface
personalization. The user model structure and
dimensions must be determined and initialized in the
set up process to avoid cold start problem, but it
must be flexible to adapt to ubiquitous changing
environment. A process of concept alignment for
interoperability will provide articulation between
heterogeneous sources. This process will
automatically maintain the ontology structure and
dimensions. We briefly describe this process in
3.2.3.
3.2 User Profile Structure and
Dimensions
People interact with many systems and applications
through various devices that gather information of
the user in order to provide a more personalized
service, but a person ends with many isolated user
profiles and models, tailored to the application,
device or system for which it was intended. The
domain of user profile is therefore very complex, but
our user model must be domain-independent. The
task of building a flexible ubiquitous user model
ontology that represents knowledge of the user of all
possible profile providers and consumers actually
used by a person and to come is very ambitious.
Many attempts have been made in the user modeling
community to provide cross-system personalization
and foster sharing and reusing user models by
developing a commonly agreed ontology of a
domain as we described in section 2, but including
in it all possible applications and context is not a
feasible solution for sharing and reusing.
We have to compromise and provide the
ontology set up with a user representation of a
flexible user profile for some commonly used
applications and devices and enable the ontology to
continually adapt to new systems and applications
requirements.
3.2.1 Ontology Set Up
The main idea of our approach is to crumble the
information extracted from supplier’s profiles at
concept granularity, and find the semantic relations
between each of them and the most similar concept
already present in the U2MIO ontology. Once the
relation is found, a link is established via semantic
relation and mapping properties. A data model that
provides us a common understanding is used in
combination with OWL as a formal knowledge
representation language. SKOS is a W3C
recommendation since 2009.
With the purpose of defining the profile structure
and dimensions considered for our ubiquitous user
model, we reviewed some of the international
initiatives towards standardization of user profile
structure, and user profile schemas of most
commonly used applications and devices.
From European Telecommunications Standards
Institute (ETSI) (ETSI, 2005), 3rd Generation
Partnership Project (3GPP) GUP (3GPP GUP, 2009)
and we established that the profile should be flexible
and redundancy must be avoided.
In a user-centric multi-application
environment, the nature of user’s profile data is
static, semi-static and dynamic. This information is
recollected explicitly or implicitly by profile
suppliers. We selected the following types of profile
suppliers and consumers in order to include
information we believe is most commonly available
in the providers’ systems and needed by profile
consumers.Therefore, for the ontology set up, we
decided that the user profile structure must contain
concepts to define knowledge in the following
dimensions.
The ubiquitous user model (U2M) is modeled as
a skos:ConceptScheme and each dimension
corresponds to a skos:Collection (Figure 1), and thus
it is modeled as an OWL class. The collection is
meant to group concepts that belong to each
dimension. In the set up process, we determined a
TowardsanOntologyforUbiquitousUserModelingInteroperability
241
Figure 1: Ubiquitous user model collections.
set of concepts to initialize the U2M concept scheme
and each collection (and sub collection) based on the
profile suppliers and consumers chosen. First of all
we decided to include FOAF ´s concepts as it is
currently considered as one of the best populated
ontologies and is extensively adopted to describe a
user and his relations (Ding et al., 2005). “The
FOAF ontology is the most widely used domain
ontology on the semantic web” (Finin et al., 2005).
We also included concepts of social network user
basic profile since these applications have a great
popularity among internet users. According to
(Nielsen, 2011) social networks continue to
dominate American’s time online. Although social
network user profile is available through dedicated
APIs, they do not always provide accurate schemas
and documentation. Efforts have been made to semi-
automatically derive social network schemas from
social network data (Kapsammer et al., 2012). So
taking into account schemas provided by Facebook,
LinkedIn, Twitter, and Google+, and generated
models and diagrams published by (Kapsammer et
al., 2012), the concept schemes and concepts were
created using Protégé.
First a concept scheme was included for each
profile provider or consumer with their concepts
related to relevant user profile data. Subsequently a
concept is included in the pertinent collection in the
U2M concept scheme avoiding redundancy.
Mappings are established via SKOS semantic
relation object properties using only
skos:exactmatch if the concepts are interchangeable.
An example of the concept Age set up is shown in
figure 2. We finally add concept schemes related to
Microsoft Health Vault personal health record and
basic observations and measurements. Credentials
needed for authentication to both the profile supplier
web site and profile consumer, were given explicitly
by the user in order to avoid privacy issues.
3.2.2 Ontology Manipulation
The ubiquitous user modeling interoperability
ontology (U2MIO) is part of an adaptive Web-based
system. The ontology is stored in OWL/XML format
and is manipulated by a C# program which includes
OwlDotNetApi (Pellens, 2006) created to
manipulate OWL ontologies. In order to integrate a
ubiquitous user model gathering information from
heterogeneous sources and be able to reuse this
information, alignment between similar concepts
must be done.
Figure 2: Example of concept set up (AGE).
3.2.3 Process of Concept Alignment for
Interoperability in Heterogeneous User
Models
The alignment process achieves interoperability
between a user model document written or translated
to XML (named source) with the ubiquitous user
model (named target). The ubiquitous user model
(u2m) will provide articulation between
heterogeneous profile suppliers/consumers given the
mappings from all individual concepts of the
document sources with the corresponding ubiquitous
user model concepts. The ubiquitous user model can
also provide articulation with profile consumers, for
example, Web Services’ preferred values when they
are requested.
We describe the process of alignment between a
given concept c
s
of the source concept scheme X
s
and the best suited for alignment skos:Concept c
tb
in
KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
242
the target concept scheme X
t
.We consider a concept
scheme as (C, H
C,
V
C
) where C is a set of concepts
arranged in a subsumption hierarchy H
C.
V
C
is the set
of corresponding concept values available.
Each concept c
s
in a set of concept source C
S
is
defined by a label string1 and subclass relationships.
The attribute or element identifier corresponds to
string1 and it is typically described as a simple or
compound word in natural language. The
hierarchical structure must be described as a
skos:ConceptScheme.
A concept on the target side C
T
is described by
set of labels included in the target skos:Concept
consisting of lexical labelling, notation and
documentation SKOS properties.
The alignment process consists of two steps:
a) An element level matching in which concepts
are directly compared to each other without
considering the hierarchy structure and values.
Its goal is given concept c
s
of the source concept
scheme X
s
, find the best concept label c
tb
from a
set of concept candidates for alignment in the
target concept scheme X
T
.
b) A structure level matching in which context of
the source and target concepts are considered.
The ultimate goal of this process is determine
one-to-one mappings between the concept c
s
of
the source concept scheme X
s
and the best
concept c
tb
from the set of labels C
t
of the target
concept scheme X
T
.
The first step in the element level matching
process is finding a set of concept candidates for
alignment in the target concept scheme X
T
. In order
to perform this task a method combining three
different similarity methods based on Dice
coefficient similarity (Dice,1945), longest common
substring and a similarity based in Wordnet lexical
base for English language (Fellbaum,2005) is
proposed. Each relation between c
s
and c
t
is
classified as equivalent, related or independent based
on previous results.
The process of structure level matching consists
of three steps:
1. Select the sets of neighbours N
s
and N
T
from
the source hierarchy H
s
and the target hierarchy
H
t
respectively, which define the context of
each concept source c
s
.
2. Calculate a new similarity matrix between the
two sets of neighbours N
s
and N
T
.
3. Applying a set of IF THEN rules to determine
the one-to-one mappings between the concept
c
s
of the source concept scheme X
s
and the best
concept c
tb
from the set of labels C
t
of the
target concept scheme X
T
.
4 DISCUSSION
In set up process, U2MIO ontology was initialized
with the core information to represent a generic
ubiquitous user model. The user-adaptive Web based
system will manipulate the ontology allowing the
user modeling interoperability. The process of
concept alignment enables the interoperability
between individual concepts from a source
document and the skos:concepts of the ubiquitous
user model ontology.
Our approach differ from standardization-based
user modeling ontologies like GUMO (Heckmann,
2005), in which user modeling interoperability
participants have to adhere to a standard. Our
ontology evolves to adapt itself to new sources and
consumers. Systems based on user profile
management selection like (Sutterer, 2009)
aggregate different versions of the user model and
automatically select the best suited version for the
user state. This approach implies redundancy and
inconsistencies between the user models of
heterogeneous sources. As we combine the concepts
of the heterogeneous sources in one user model
represented by the ontology, the system resolves the
conflicts between the semantic meanings avoiding
redundancy and inconsistencies. In other mediation
based user modeling systems (Van der Sluijs &
Houben, 2006) mappings are made by human effort
making it hard to include new participants in user
modeling interoperability. In (Carmagnola,2009) the
authors presents a highly flexible system, but to take
part in the interoperability process, every provider
need to comply to a standard format for the
exchange and maintain a sharable user model which
includes the fragments of user model as RDF
statements. User modeling systems must do the
effort of maintaining a sharable user model in a
standard format. In our approach, the effort of
syntactic and semantic interoperability is made by
the user-adaptive system.
5 CONCLUSIONS
We present U2MIO ontology for user modeling
interoperability based in SKOS ontology. The
ontology initially includes the core dimensions and
profile structure to represent a generic ubiquitous
user model, and the user-adaptive system provides
the possibility for the user model to evolve during
the time it is used. The U2MIO provides mediation
based user modeling for sharing and reusing
information from heterogeneous user models. A
TowardsanOntologyforUbiquitousUserModelingInteroperability
243
two-tier matching strategy is proposed for the
process of concept alignment that permits the
interoperability between profile suppliers and
consumers .Our on-going and future work includes
further research in the automatic transformations
needed at instance level in order to establish
interchange ability at instance level.
REFERENCES
3rd Generation Partnership Project; Technical
Specification Group Services and System Aspects;
3GPP Generic User Profile (GUP); Architecture Stage
2, Release 9, 3GPP TS 23.240 V9.0.0, 2012, Available
http://www.3gpp.org/ftp/Specs/html-info/23240.htm.
Berkovsky, S., Kuflik, T., Ricci F., 2007, Mediation of
user models for enhanced personalization in
recommender systems. User Modeling and User
Adapted Interaction.
Brickley, D.; Miller, L.; FOAF vocabulary specification
0.91 Online at: http://xmlns. com/foaf/spec, 2007.
Carmagnola, F. and Cena, F., 2009, “User identification
for cross-system personalisation,” Journal of
Information Science, vol. 179, no. 1-2,pp. 16–32.
Carmagnola, F., 2009, Handling Semantic Heterogeneity
in Interoperable Distributed User Models, Advances in
Ubiquitous User Modelling: 20-36.
Dice L. R. (1945). Measures of the amount of ecologic
association between species. Ecology 26: 297-30
Ding, L., Zhou, L., Finin, T. and Joshi, A., 2005, “How
the semantic web is being used: an analysis of FOAF”,
Proceedings of the 38th International Conference on
System Sciences, Hawaii, January.
ETSI EG 202 325, Human Factors (HF), User Profile
Management, 2005.
Fellbaum, Christiane (2005). WordNet and wordnets. In:
Brown, Keith et al. (eds.), Encyclopedia of Language
and Linguistics, Second Edition, Oxford: Elsevier,
665-670
Finin, T., Ding, L. Zhou, L., Joshi, A., 2005,"Social
networking on the semantic web", The Learning
Organization, Vol. 12 Iss: 5, pp.418 – 435.
Heckmann, D., 2005, Ubiquitous User Modeling, PhD
thesis, Department of Computer Science, Saarland
University, Germany.
Houben, G. et al., 2005, “State of art: semantic
interoperability for distributed user profiles”,
Telematica Institute Report, Project Alter Ego/D1.6 v
2, 2.
Iqbal, Z., Noll, J., Alam, S., “Role of User Profile in
Cloud-based Collaboration Services for Innovation”
International Journal on Advances in Security, vol 4
no 1 & 2, year 2011, http://www.iariajournals.org/
security/
Iqbal, Z., Noll, J., Alam, S., Chowdhury, M.,13-19 June
2010, “Toward User-centric Privacy-aware User
Profile Ontology for Future Services”, In the
Proceedings of IEEE Third International Conference
on Communication Theory, Reliability, and Quality of
Service (CTRQ 2010), Athens, Greece.
Kapsammer, E. et al., 2012, User Profile Integration Made
Easy - Model-Driven Extraction and Transformation
of Social Network Schemas, Paper at International
Workshop on Interoperability of User Profiles in
Multi-Application Web Environments, at 21st
International World Wide Web Conference (WWW),
Lyon, France, April 16 - 20, ACM Publications.
Leonardi, E., Houben G-J, van der Sluijs, K,, Hidders, J.,
Herder E., Abel F, Krause D, Heckmann, D:
June,2009, User Profile Elicitation and Conversion in
a Mashup Environment. In Int. Workshop on
Lightweight Integration on the Web, in conjunction
with ICWE 09, San Sebastian, Spain.
Miles, A and Pérez-Agüera, J., 2007, SKOS: Simple
Knowledge Organization for the Web, Cataloging &
Classification Quarterly, Vol. 43, No.3.2007, pp. 69-
83.
Nielsen: Social Media Report: Q3 2011, Available in
http://blog.nielsen.com/nielsenwire/social/, consulted:
April 2012.
Pellens, B., OwlDotNetApi, submitted in 2006, Available:
http://users.skynet.be/bpellens/OwlDotNetApi/owldot
netapi.html
Sutterer, M. “User Profile Management and Selection in
Context-Aware Service Platforms for Networks
Beyond 3G: A Practical Application of Semantic Web
Technologies”, October 2009, PhD Dissertation,
Universität Kassel.
Van Der Sluijs, K., Houben, G.,2006, A generic
component for exchanging user models between web-
based systems, In International Journal of Continuing
Engineering Education and Life-Long Learning,
2006,Vol.16,No.1/2 pp. 64-76.
Viviani, M., Bennani, N., Egyed-zsigmond, E. and Liris,
I., 2010, A Survey on User Modeling in Multi-
Application Environments. In Third International
Conference on Advances in Human-Oriented and
Personalized Mechanisms, Technologies and Services.
KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment
244
