MANAGING USER INTERACTION IN AN
ONTOLOGY-BASED SYSTEM
Anna Goy and Diego Magro
Dipartimento di Informatica, Università di Torino, C.Svizzera 185, Torino, Italy
Keywords: User interaction, User interfaces, Ontology, Semantic Web, e-Business.
Abstract: In this paper, we present an approach to the management of user interaction within an ontology-based
system. In particular, we discuss the case of ARNEIS, an “intelligent” web-based repository of software
solutions, that enables software houses to upload a description of their software solutions for business
automation, and small-to-medium sized enterprises to find software products or services supporting their
activities. To this purpose, as argued in the paper, a particularly meaningful field is Customer Relationship
Management, that is thus exploited as a test bed domain. The functionality of ARNEIS is based on an
ontological representation of the domain knowledge, which represents the shared conceptual vocabulary to
express software descriptions and technological requirements. In this paper we describe our proposal for the
management of a user-friendly interaction enabling software houses to upload the semantic representation of
the description of their offers. The approach we propose within the ARNEIS scenario represents an example
of a more general solution to face the issue of how to build formal representations of resources in an
ontology-based IR system.
1 INTRODUCTION
In order to be competitive in the globalized market
Small-to-Medium sized Enterprises (SME) have to
exploit the support provided by innovative
technological solutions to their business. On their
side, Information and Communication Technologies
(ICT) relies more and more on Internet and Web
technologies: web-based solutions are available for
almost any kind of application, supported by
innovative technologies at different levels, ranging
from Cloud Computing infrastructures (Creeger,
2009; Dikaiakos et al., 2009), to Web Services
(Alonso et al, 2004), to Semantic Web (Antoniou
and Van Harmelen, 2004).
However, the technological support is not
enough: in order to be competitive, SME should also
take care of new business and management
approaches. An example is the way in which they
handle their relationships with customers. The new
market requires personalized approaches to the
single customer, as well as flexible offers, that need
to be updated rapidly. Moreover, in order to be
aware of the market and customer behavior trends,
data about offers, sales, communications with
customers, and so on have to be elaborated very
rapidly, to support management and marketing deci-
sions. For these reasons, SME pay every day more
and more attention to the principles of Customer
Relationship Management (Freeland, 2005) and to
those ICT products and services supporting it. The
key feature of the CRM approach is a one-to-one
marketing perspective, i.e. establishing personalized
relationships with the single customer, by producing
personalized offers, pricing, after-sale services, etc.
Moreover, CRM is a field in which technological
innovation could bring great benefits since it
requires the processing, integration, and analysis of a
huge amount of heterogeneous knowledge (about
customers, sales, communications, etc.); it also
requires effective, fast and integrated
communication tools.
Within this scenario, we think that there are two
main issues that play a major role:
(1) A clean, complete, and sharable ontology
(Guarino, 1997) defining the concepts related to
CRM is of paramount importance, in order to
support the mentioned knowledge management
activities, as well as Enterprise Application
Integration; see, for instance, (Benjamins, 2008).
(2) SME would get great benefits from a web-based
service supporting an intelligent matching between
supply and demand for CRM-related tools.
194
Goy A. and Magro D..
MANAGING USER INTERACTION IN AN ONTOLOGY-BASED SYSTEM .
DOI: 10.5220/0003296401940201
In Proceedings of the 7th International Conference on Web Information Systems and Technologies (WEBIST-2011), pages 194-201
ISBN: 978-989-8425-51-5
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
Moreover, the shared knowledge mentioned in (1)
can be exploited in a cross-enterprise scenario like
the one outlined in (2). Thus, we designed an
architecture for a web-based intelligent system
supporting SME in finding suitable software
solutions for their business, and we chose CRM as a
testbed for this architecture.
On the basis of this architecture, we developed
ARNEIS (Advanced Repository for Needs of
Enterprises and Innovative Software), a prototype
implementation of a web-based repository of
software solutions, that exploits Web Services and
Semantic Web technologies. ARNEIS users are: (1)
ICT companies (i.e., software houses) that offer
software solutions for business automation, but can
be in trouble in getting in contact with their potential
customers; (2) SME that aim at finding software
products or services supporting their business and
management activities, but lack the know-how to
find the right ICT solution that fits their needs.
Such an intelligent repository has to be equipped
with a large and detailed knowledge base, e.g. an
ontology, which represents the shared conceptual
vocabulary; such an ontology is the basis for the
matching algorithm used to suggest SME the most
suitable software solutions, given their needs.
Moreover, ARNEIS requires a web-based user
interface (UI) enabling ICT companies to describe
their software solutions and SME to express their
requirements and needs and to find suitable software
solutions supporting their business.
We claim that both these goals, i.e. building the
CRM ontology and defining the system UI, have to
be based on a domain analysis that takes into
account how users, both from ICT companies and
from SME, talk (and think) about CRM activities
(see Section 3).
The CRM ontology developed for the ARNEIS
project is described in (Magro and Goy, 2008a;
Magro and Goy, 2008b), while this paper focuses on
the definition of the UI enabling the users to interact
with it. In particular, Section 2 briefly describes the
system architecture and Section 3 reports the main
results of the domain analysis; Section 4 focus on
the UI management: it provides a brief survey of the
relevant related work in the field, it discusses the UI
design choices and it describes the mechanism for
UI management. Section 5 briefly discusses some
open issues and concludes the paper.
2 ARNEIS ARCHITECTURE
The basic architecture of the ARNEIS system is des-
cribed in (Goy et al., 2008); Figure 1 shows a
simplified version of it.
Figure 1: ARNEIS system architecture (a simplified
version taken from (Goy et al., 2008c)).
It is a standard three-tiers architecture, where the
presentation layer is represented by a standard web
browser.
The application logic is contained in the ARNEIS
core, which includes three main components: the UI
Manager, the Knowledge Manager, and the
Matching Engine.
The UI Manager generates the UI taking as input
the concepts, properties and relations provided by
the Knowledge Manager and represented in the
Ontology; moreover, it collects information
provided by the user and forwards it to the
Knowledge Manager, that builds the semantic
representation of software descriptions and SME
requirements.
The Knowledge Manager mediates the
interaction between the UI Manager and the data
layer (its role will be discussed in more detail in
Section 4.3). Moreover, it dialogs with the Matching
Engine (whose description is out of the scope of this
paper), by providing it the semantic representations
on the basis of which the Matching Engine will
calculate the correspondence between SME
requirements and software house offers.
The data layer contains the Knowledge Base,
whose main components are the Ontology,
representing the system semantic competence about
the domain (CRM), and the semantic representation
of software descriptions and SME requirements
(Semantic Descr KB). Both the Ontology and the
Semantic Descr KB are written in OWL
(http://www.w3.org/2004/OWL/).
The OWL domain Ontology is based on a CRM
Reference Ontology (Guarino, 1997), described in
(Magro and Goy, 2008a; Magro and Goy, 2008b),
which in turn is based on DOLCE (http://www.loa-
cnr.it/DOLCE.html). The CRM Reference Ontology
basically models: (1) business activities (e.g., sales,
offers, communications, appointments, etc.); (2)
business relationships which the company is
involved in (e.g., relationships with actual or
potential customers); (3) the knowledge that the
MANAGING USER INTERACTION IN AN ONTOLOGY-BASED SYSTEM
195
company has on (or derives from) business activities
and relationships; (4) software supporting business
activities and knowledge management.
3 DOMAIN ANALYSIS
Given the architecture described in the previous
section, which is the information needed by the UI
Manager to generate the UI? And which kind of
information should be elicited from the user in order
to build a structured and formal representation of the
software solutions offered?
To answer this question, we carried out a
detailed domain analysis, in order to understand how
users (both software houses and SME) talk about
CRM.
We analyzed two types of information sources:
(1) Documents (e.g., brochures, white papers),
produced by ICT companies, describing their CRM
software tools. (2) Interviews with (a) salesmen
from ICT companies, aimed at eliciting the way they
describe software solutions for CRM for SME; (b)
managers of SME, in order to understand which
concepts and terms they use to think about their
activities related to customer management.
The analysis of documents and interviews
provided us with the following outcomes:
(1) We identified the main concepts, properties and
relations involved in the description of CRM
activities in SME, which represent the basic
requirements for building the CRM Ontology.
(2) We identified a set of dialog topics, i.e. concepts
that emerged to be the keys of the description of
CRM activities. Within the system, these concepts
can be represented as templates, i.e. general
conceptual patterns that are instantiated with more
specific concepts in the different
documents/interviews, by means of various
linguistic forms. Each template has a corresponding
formal semantic representation in terms of the
Ontology. Basically, each description of a software
solution (or a SME requirement) is an instantiation
of a number of such templates (see Section 4.3).
4 USER INTERACTION
As mentioned in Section 2, the system needs a
formal representation of the descriptions of software
solutions and of technological requirements (both
based on the terms defined in the system Ontology)
in order to find the software solutions best matching
the SME requirements. Since it is unrealistic to ask
users to write descriptions in a formal Semantic Web
language, such as OWL, the generation of a user-
friendly UI turned out to be an issue of major
importance.
4.1 Related Work
The importance of the UI to access complex
knowledge bases is a topic largely studied in various
research fields. Many authors recognize that,
recently, there has been an increase of interest in the
design of UI for systems based on semantic
technologies (see, for instance, the SWUI workshops
series). The main issue within this field is, quite
obviously, the fact that the user interacting with a
formally encoded knowledge base (e.g., an
ontology) should not be exposed to the details of
such a formal representation. In other words,
“semantic technologies should be invisible to users”
(Benjamins, 2008), p. 76.
The most traditional approaches facing this issue
are devoted to support user queries aimed at
searching for information encoded in formal
knowledge bases (databases or ontologies). These
approaches are mainly based on the translation of
keywords (provided by the users) into formal queries
(being SQL, or DL, or other formal languages); see,
for instance, (Tran et al., 2007b).
Among many approaches to build UI to access
knowledge bases, some studies propose to use NL-
based UI; e.g. (Bernstein and Kaufmann, 2006),
(Cimiano et al., 2008), (Damljanovic et al., 2010).
Other approaches propose graphical (web-based) UI;
e.g., (Thoméré et al., 2002).
(Bhagdev et al., 2008a) and (Bhagdev et al.,
2008b) describe two tools supporting the user
interaction with formal knowledge bases: (a) K-
Forms is a tool enabling users to define knowledge
structure and content through a user-friendly form-
based UI; the tool then translate such a knowledge
into a formal (RDF/OWL) representation; (b) K-
Search is a tool that supports knowledge search and
sharing.
In all the mentioned approaches the user input is
a query aimed at extracting information from a
(formal) knowledge base. In ARNEIS, the two types
of users interacting with the system have different
goals: the goal of a user from a software house is to
provide the system with a description of a software
solution, hoping that it will match some user
requirements; the goal of a user from a SME is to
“explain” the system the SME technological needs
in order to find a suitable software solution. Thus, in
WEBIST 2011 - 7th International Conference on Web Information Systems and Technologies
196
ARNEIS, in both cases, the user input is a complex
description (of a software product or of a set of
requirements) based on an existing ontology. In
order to elicit this kind of information the system
requires a UI that is more complex than the UI
typically provided by the previously mentioned
approaches, where the user simply provides the
system with a query.
Other works concerned with user interaction with
formal representations are authoring tools: in these
cases the UI is aimed at enabling the user to design
and populate a complex knowledge structure such as
an ontology (e.g. Protégé:
http://protege.stanford.edu/). With respect to these
tools, ARNEIS has, not only a different goal, but
also different types of users: in fact, ARNEIS users
are not expert in knowledge representation and they
could find it difficult to interact with an authoring
system, even if provided with a smart (maybe
graphical) UI.
Some approaches to Semantic Search propose to
build ontology-based Information Retrieval (IR)
systems, i.e. systems in which both resources and
user queries are represented in a formal semantic
language. For example (Tran et al., 2007a) proposes
a resource model based on various integrated
ontologies, expressed in OWL, enabling resources
(documents) representation in terms of ontology
elements (i.e., entities and axioms). As discussed by
the authors, the issue of how to obtain formal se-
mantic representations of resources is still open.
Some steps have been made towards the automatic
extraction of such representations from (textual)
documents (e.g., Banko et al, 2007), but the results
in this field are still poor. As an alternative, “a
manual approach can be undertaken” (Tran et al.,
2007a), p. 65. In particular, we think that a manual
approach actually makes sense in those cases in
which resources are not already available as
documents. For instance, in a scenario as the one
described above, a software house may have only
simple brochures describing its software products; it
seems to be reasonable to provide it with a tool
supporting the construction of a formal semantic
representation of their software products, which can
be handled as resources by an ontology-based IR
system. Our proposal is as a step in this direction.
In the following section we will describe how we
faced the issue of the design of the user interaction
in ARNEIS. However, the proposed approach aims
at providing a general solution to solve the problem
of how to build formal representations of resources
in an ontology-based IR system.
4.2 User Interface
Concerning the type of UI most suited to our system,
we chose on-line forms, since they are very familiar
to web users; see (Bhagdev et al., 2008a). However,
usually, users from SME, interacting with ARNEIS
Figure 2: “(Dynamic) acquisition of data from (another) enterprise application” template.
MANAGING USER INTERACTION IN AN ONTOLOGY-BASED SYSTEM
197
in order to look for software solutions supporting
their CRM activities, are not technical experts and
thus they can be in trouble even in filling in large
on-line forms, possibly requiring technical skills in
order to be completed. Thus, we are studying the
possibility of implementing different types of UI for
these users: a Natural Language (NL) UI, in which
the SME users can freely express their requirements,
and a UI based on a graphical representation of
business processes (BP), based on the idea that
business processes could be the form in which
companies think about their business and
management activities.
These two alternative UIs are a work in progress.
The form-based UI, instead, can be suited for ICT
company users, which are probably skilled enough
about the functional and technological aspects of
their software products or services. Moreover, such
users are probably also more used to web-based
interaction, often consisting in a sequence of on-line
forms. Finally, such users are probably also more
motivated to complete a possibly boring task, since
they are describing the software solutions they offer,
and this can be viewed as an effective promotion of
their products and services.
Thus, in the following, we will concentrate on
the form-based UI devoted to the acquisition of
descriptions of software solutions by ICT
companies.
4.3 Managing User Interaction
We claim that the design of the user interaction
enabling ICT company users to provide a description
of their software solutions could only be based on an
analysis of the way in which users talk about CRM.
For this reason, in the domain analysis phase,
besides Ontology requirements, we also identified
dialog topics, representing the key concepts of the
descriptions of software solution supporting CRM
activities (see Section 3).
In order to clarify how dialog topics are
exploited in the system, let’s look at an example.
One of the templates representing dialog topics we
extracted from our analysis corresponds to the
concept of “(dynamic) acquisition of data from
(another) enterprise application”, which means that
the described CRM software application can acquire
data by directly communicating with another
application (within the system, templates do not
have a linguistic form, but only a formal, OWL, one.
For the sake of readability, here we provide also a
rough linguistic “translation”). Such a general
concept is instantiated in different
documents/interviews with more specific concepts
by heterogeneous linguistic expressions: specific
instances of this template include more specific
concepts in place of “data” and in place of
“enterprise application”; for example “real time
acquisition of customers info from Ms Outlook”, or
“(management of) product records acquired from
ERP software”.
We decided to represent templates as ontological
concepts belonging to an Application Ontology
(Guarino, 1997), specific for ARNEIS, and linked to
the upper CRM Ontology.
A template represents a structured concept in
which there are some variables (slots). By default,
such variables are filled in by generic concepts (e.g.
“data”), that can be replaced by more specific
concepts (e.g. “product data”).
Figure 2 shows the OWL logical representation
(generated by Protégé) of the class representing the
template corresponding to the concept of “(dynamic)
acquisition of data from (another) enterprise
application”.
In order to identify slots within a template, we
needed a way lo “label” them, possibly without
compromising the correctness of the OWL syntax.
Thus, we decided to indicate them by adding, to
each concept that represents a default slot filler, the
expression and S
i
, where S
i
is a unique identifier
(automatically generated) for that slot. Thus, for
example, in Figure 2 the expression (CRM_element
and S
1
) identifies a slot (labeled S
1
) filled in, by
default, by the CRM_element class (intuitively
speaking, CRM_element represents all the items that
are typically involved in CRM: customers, products,
orders, sales, and so on). If we aim at expressing the
acquisition of data about customers (e.g., “customers
info”), in the template instantiation, the class
CRM_element will be replaced by a more specific
one, i.e. Customer (subclass of CRM_element) and
the expression and S
1
will be deleted; if we aim at
expressing the acquisition of data about products
(e.g., “product records”), CRM_element will be
replaced by Product_or_service (subclass of
CRM_element). Analogously, the class
(Application_software and S
2
) can be replaced by
the class Ms_Outlook, or ERP.
At a first glance, this labeling solution may seem
to bring to odd meanings. In fact, from a semantic
point of view, by writing Information_object ...
refers_to some (CRM_element and S
1
), we are
actually saying that an Information_object refers to
something that is, at the same time, a CRM_element
and a S
1
(see Figure 2), which is not the meaning we
would like to express. Thus, it is important to stress
WEBIST 2011 - 7th International Conference on Web Information Systems and Technologies
198
that the expression and S
i
is used merely as a label,
to identify slots within a template. We chose this
solution because it preserves the syntactic
correctness of the OWL representation. Moreover, it
does not produce odd semantic results, since in
ARNEIS not instantiated templates are never used in
any form of semantic reasoning.
For each template, the UI Manager has to
generate a set of web-forms, aimed at eliciting the
information needed for filling in the slots. Since,
typically, the knowledge on the basis of which such
forms are generated does not change in time, web
forms are pre-compiled off-line; they are re-
generated only in case the knowledge bases are
modified.
In order to enable the UI Manager to generate the
web forms, the Knowledge Manager performs two
(nested) steps:
(1) For each template in the Application Ontology, it
extracts all the slots by looking for the expressions
of the form (C and S
i
), where C is a named class of
the reference Ontology.
(2) For each slot (identified by S
i
), the Knowledge
Manager extracts from the Ontology all the
subclasses of C and provides the UI Manager with
this information.
The UI Manager, in turn, for each slot, generates a
web form in order to ask the user which subclass of
C (if any) she wants to consider.
For example, Figure 3 shows the form asking the
user to select the subclasses of CRM_element.
Figure 3: web form for filling in the first slot of the
template corresponding to “(dynamic) acquisition of data
from (another) enterprise application”.
The user selects the desired subclass (C
x
) and the
UI Manager sends this information back to the
Knowledge Manager, that substitutes C
x
in place of
(C and S
i
) within the corresponding template slot.
The result (that refers to the example presented
above) is shown in Figure 4.
Figure 4: part of the instantiated template after the user
answer.
There are some issues to be faced in order to
make this mechanism work: (1) Which is the proper
question the UI Manager should pose to the user
(e.g., in Figure 3, “Would you like to acquire data
concerning...”)? (2) In some cases, the user may
want to specify a more specific class (e.g., she wants
to acquire data about golden customers, which are
represented, within the Ontology, by the class
Golden_customer, subclass of Customer). (3) It is
quite common that the suited value for filling in a
slot depends on the value assigned to another slot (of
the same template). In the example above (see
Figure 2), if we fill in the slot S
1
with Customer,
meaning that we want to acquire information about
customers, the slot S
5
(representing the knowledge
structure that is modified by the data acquisition)
have to be filled in with Customer_database.
Moreover, many slots within a template are usually
filled in by the same concept. For instance, in Figure
2, if we fill in the slot S
1
with Customer, the slots S
7
and S
11
have to be filled in with Customer too.
In the following we will explain how we faced
the mentioned issues.
(1) Natural Language Questions. The knowledge
engineer who configures ARNEIS on a new domain
is in charge of defining the Application Ontology
containing the templates. When doing this, she
identifies the slots and, for each slot, she provides a
natural language question that will be used by the UI
Manager to generate the form referring to that slot.
The link between the slot S
i
within the template T
j
(T
j
.S
i
) and the corresponding question is stored in a
Configuration KB which is accessed by the UI
Manager during the form generation process.
(2) Indirect Subclasses. The Knowledge Manager,
when extracting from the Ontology all the subclasses
of C (step (2) mentioned above), actually extracts
not only the direct subclasses of C, but the whole
subtree. The UI Manager, on the basis of this
information, includes a link to enable the user to
optionally expand the subtree (see “view more” links
in Figure 3): in this way she is enabled to select any
(direct or indirect) subclass she is interested in.
However, given the complexity of the Ontology,
listing all the (named) subclasses of a given class C
could result in a too long list, containing concepts
that are relevant from the formal point of view, but
not from the user perspective. For this reason, we
added to the Configuration KB, for each slot (C and
S
i
), identified by the unique label T
j
.S
i
, a list of the
subclasses of C that should be asked to the user. For
example (the Configuration KB is an XML
document; here we show its content in pseudo-code
instead of XML for the sake of readability):
MANAGING USER INTERACTION IN AN ONTOLOGY-BASED SYSTEM
199
T
3
.S
1
{Sales_agent, Customer, ...}
The list of “subclasses to be asked” potentially
contains both direct and indirect subclasses of C.
(3) Dependencies between slots. In order to take into
account possible dependencies between slot fillers,
the Configuration KB also contains some if-then
rules representing such dependencies; for instance:
IF T
3
.S
1
= Customer
THEN T
3
.S
5
= Customer_database
IF T
3
.S
1
= Customer
THEN T
3
.S
7
= Customer
These dependency rules enable the UI Manager to
avoid asking the user useless questions: after having
asked the filler for slot S
1
, slots S
5
and S
7
will be
filled in automatically.
When the user has completed all the forms
proposed by the ARNEIS system, the instantiated
templates are saved in the Semantic Descr KB (see
Section 2): the set of such instantiated templates is
the OWL representation of the software solution
supporting CRM, proposed by the ICT company.
5 DISCUSSION
AND CONCLUSIONS
In this paper, we described the management of the
user interaction within the ARNEIS system, an
“intelligent” web-based repository of software
solutions, that enables ICT companies to upload a
description of their software solutions for business
automation, and SME to find software products or
services supporting their business and management
activities. The functionality of such an intelligent
repository is based on a semantic representation of
the domain knowledge, which represents the shared
vocabulary to express both software descriptions and
technological requirements. In particular, we
described the management of a user-friendly
interaction enabling software houses to upload the
semantic representation of the description of their
products and services. The approach we proposed
within the ARNEIS scenario suggests a general
solution to solve the problem of how to build formal
representations of resources in an ontology-based IR
system.
Our approach is strongly based on a domain
analysis that takes into account how users talk about
their business activities and the software
applications that could support such activities. In
fact, the mechanism described in this paper relies on
templates stored in an Application Ontology and a
Configuration KB, that have to be built by the
knowledge engineer who configures the system on a
domain (e.g., CRM). This task could represent a
considerable knowledge acquisition effort. However,
the definition of templates and Configuration KB is
typically done once and usually does not require
frequent updates. Moreover, the Configuration KB
definition can be easily supported by a user-friendly
tool that provides the knowledge engineer with
simple mechanisms supported by system defaults
(e.g., the lists of the subclasses of C that have to be
proposed to the user can be automatically built by
default including all the available subclasses; the
knowledge engineer will be enabled to simply delete
those ones she thinks are not meaningful for the end
user).
Since the knowledge acquisition effort is a key
aspect of knowledge-based systems like ARNEIS,
the development of the mentioned user-friendly tool,
supporting the knowledge engineer in the system
configuration on new domains, represents our main
future work.
REFERENCES
Alonso, G., Casati, F., Kuno, H., Machiraju, V., 2004.
Web Services. Springer, Berlin.
Antoniou, G., Van Harmelen, F., 2004. A Semantic Web
Primer. The MIT Press, Cambridge MA.
Banko, M., Cafarella, M. J., Soderland S., Broadhead M.,
Etzioni, O., 2007. Open information extraction from
the web. In IJCAI 2007, 20th International Joint
Conference on Artificial Intelligence, 2670-2676.
Morgan Kaufmann.
Benjamins, V. R., 2008. Near-Term Prospects for
semantic Technologies. In IEEE Intelligent Systems,
23(1), 76-88.
Bernstein, A., Kaufmann, E., 2006. GINO - A Guided
Input Natural Lanaguage Ontology Editor. In ISWC
2006, 4th International Semantic Web Conference,
114–157. Springer.
Bhagdev, R., Chakravarthy, A., Chapman, S., Ciravegna,
F., Lanfranchi, V., 2008a. Creating and Using
Organisational Semantic Webs in Large Networked
Organisations. In ISWC 2008, 6th International
Semantic Web Conference, 723-736. Springer.
Bhagdev, R., Chapman, S., Ciravegna, F., Lanfranchi, V.,
Petrelli, D., 2008b. Hybrid Search: Effectively
combining Keywords and Semantic Search. In ESWC
2008, XX European Semantic Web Conference, 554-
568. Springer.
Magro, D., Goy, A., 2008a. The Business Knowledge for
Customer Relationship Management: an Ontological
Perspective. In Int. ACM Workshop on Ontology-
supported business intelligence. ACM Press.
WEBIST 2011 - 7th International Conference on Web Information Systems and Technologies
200
Magro, D., Goy, A., 2008b. Towards a first Ontology for
Customer Relationship Management. In IEEE/ACM
Workshop on Applied Ontologies in Distributed
Systems, 637-643. ACM Press.
Goy, A., Magro, D., Prato, F., 2008. ARNEIS: A Web-
based Intelligent Repository of ICT Solutions for E-
business. In Int. ACM Conference on Information
Integration and Web-based Application & Services,
403-406. ACM Press.
Cimiano, P., Haase, P., Heizmann, J., Mantel, M., 2008.
ORAKEL: A Portable Natural Language Interface to
Knowledge Bases. In Data & Knowledge Engineering,
65(2), 325–354.
Creeger, M., 2009. CTO roundtable: Cloud computing. In
Communications of the ACM, 52(8), 50–65.
Damljanovic, D., Agatonovic, M., Cunningham, H., 2010.
Natural Language Interfaces to Ontologies:
Combining Syntactic Analysis and Ontology-based
Lookup through the User Interaction. In ESWC 2010,
7th European Semantic Web Conference, 106-120.
Springer.
Dikaiakos, M. D., Pallis, G., Katsaros, D., Mehra, P.,
Vakali, A., 2009. Cloud computing. Distributed
internet computing for IT and scientific research. In
IEEE Internet Computing, 13(5), 10–13.
Freeland, J., 2005. The Ultimate CRM Handbook.
McGraw-Hill, New York.
Guarino, N., 1997. Understanding, Building and Using
Ontologies. In Int. Journal of Human-Computer
Studies, 46, 293-310.
Thoméré, J., Barker, K., Chaudhri, V., Clark, P., Eriksen,
M., Mishra, S., Portr, B., Rodriguez, A., 2002. A Web-
based Ontology Browsing and Editing System. In
AAAI 18th National Conference on Artificial
Intelligence, 927-934. AAAI Press.
Tran, T., Bloehdorn, S., Cimiano, P., Haase, P., 2007a.
Expressive Resource Descriptions for Ontology-Based
Information Retrieval. In ICTIR 2007, 1st
International Conference on the Theory of Information
Retrieval, 55-68. Springer.
Tran, T., Cimiano, P., Rudolph, S., Studer, R., 2007b.
Ontology-based Interpretation of Keywords for
Semantic Search. In ISWC 2007, 5th International
Semantic Web Conference, 523-536. Springer.
MANAGING USER INTERACTION IN AN ONTOLOGY-BASED SYSTEM
201