BUILDING E-COMMERCE WEB APPLICATIONS:

Agent- and Ontology-based Interface Adaptivity

Oscar Martinez, Federico Botella

Operations Research Center, University Miguel Hernández of Elche, Avda.Universidad, s/n, 03202 Elche, Spain

Antonio Fernández-Caballero, Pascual González

Laboratory of User Interaction and Software Engineering (LoUISE), University of Castilla-La Mancha, Campus

Universitario, 02071 Albacete, SPAIN

Keywords: E-Commerce Applications, Agent-Based Techniques, Business Architectures.

Abstract: E-Commerce Web based applications designed to facilitate Data-exchange collaboration are enjoying

growing popularity. In the next few years, business companies will want their web resources linked to

ontological content –because of the many powerful tools that will be available for using it by potential

customers. Thus, product information will be exchanged between applications, allowing computer programs

to collect and process web content, and to exchange information freely with each other. In this paper, few

pointers are used for this emerging area, and then go on to show how the ontology languages of the

semantic web can lead directly to more powerful agent-based approaches to using services offered on the

web. As a result, e-commerce architecture is outlined as an agent-based system to retrieve information

products. In this framework, an ontology representing fashion clothing domain used by potential consumers

is also introduced, where RDF-S (Resource Description Framework Schema) is used.

1 INTRODUCTION

In the next few years virtually every business

company, university or government agency would

want their web resources linked to ontological

content –because of the many powerful tools that

will be available for using it. Information will be

exchanged between applications, allowing computer

programs to collect and process web content, and to

exchange information freely with each other

(Hendler, 2001). On top of this infrastructure,

ontology negotiation between intelligent information

agents (Bailin & Truszkowski, 2001) will become

much more practical, in fact distributed computer

programs interacting with non-local web-based

resources may eventually become the dominant way

in which computers interact with humans and each

other, and will be a primary means of computation

in the not-so-distant future. Nonetheless, for this

vision to become a reality, a phenomenon similar to

the early days of the web must occur.

Concretely, e.g. in a e-commerce context, the

semantic web initiative (Berners-Lee, Hendler, &

Lassila, 2001) reflects this problem by ”giving

information a well-defined meaning, better enabling

computers and people to work in cooperation”. In

addition, Adaptive web, as envisioned in

(Brusilovsky & Maybury, 2002), should provide

business companies with optimized access to

distributed electronic product information on the

web according to particular needs of individual

consumer or group of consumers. However, the

main problem of current web systems their inability

to support different needs of individual consumer.

This can be achieved by making metadata about

different resources explicit using standardized

descriptions (SWCP, 2001).

The remainder of this paper is organized as

follows. In section 2 we present our motivation

about this emerging area like semantic resolution for

e-commerce. In section 3 we introduce our design

scenario in order to use ontologies and powerful

agent-based approaches. In section 4 we outline the

architecture of our e-commerce framework as an

agent-based system to retrieve information products.

Finally, in section 5 we present our conclusions and

future work being in progress.

351

Martinez O., Botella F., Fernández-Caballero A. and González P. (2005).

BUILDING E-COMMERCE WEB APPLICATIONS: Agent- and Ontology-based Interface Adaptivity.

In Proceedings of the First International Conference on Web Information Systems and Technologies, pages 351-354

DOI: 10.5220/0001230903510354

 SciTePress

2 MOTIVATION

Understanding the meaning of messages exchanged

between software agents has long been recognized

as a key challenge to interoperable multi-agent

systems. Several proposed solutions (Hendler 2001)

deals with for forcing all agents to use a common

vocabulary defined inside one core ontology where

semantic differences between individual agents in

the system should be allowed and be resolved when

they arise during agent interaction. For that reason,

we have started with a simple e-commerce

framework of buying and selling fashion clothing

through internet in order to represent the first step of

our ongoing effort toward a comprehensive solution

to the problem of semantic resolution.

3 FRAMEWORK DESCRIPTION

In this section we introduce a simple context of an

agent-based system to retrieve information resources

for a specific e-commerce environment. We outline

in Figure 1 this semantic web scenario.

Figure 1: Semantic E-commerce Scenario

Firstly, we have developed a simple pilot

semantic web explorer named e-fashion where users

(i.e. potential consumers) could browse and search

the information gathered in this repository. This

semantic explorer is frame-based and allows

searching metadata from fashion products; then

search results are displayed in the right frame. In the

left frame, a categorization of fashion products is

displayed, used for searching information product by

consumers.

We have proposed to build up this prototype

using the eXtended Markup Language (XML)

(XML, 1998) which has emerged in the Internet

world as a standard representation format and

because of can be useful to describe and transmit

management information. However, XML formats

alone do not give formal semantics to it. To solve

this question, we have selected an ontology language

based on the Resource Description Framework

(RDF) (RDF, 1999) and its Schema (RDF-S) (RDF-

S, 2004). Both are XML-based languages that

provide a satisfactory way of defining structured sets

of terms, with class hierarchies, and domain and

range constraints. In fact, some ontology definition

languages used in the Semantic Web are based on

RDF-S so machines can perform useful reasoning

tasks. Then, these ontology languages can be used to

improve the semantic expressiveness of the

management information specifications. With them

it is even possible to reason with the knowledge

handled in the management tasks. More to this point,

RDF-S recommendation offers terms that allow the

representation of concepts, their relationships and

their attributes, all of which formed the metadata.

By last, agents are autonomous software

components (Tveit, 2001) that possess the following

properties:

(1) autonomy: operate without intervention and

have control over their states and actions;

(2) reactivity: perceptive and are aware of their

environment and have the ability to respond in a

timely manner to the changes and actions that occur;

(3) proactiveness: take the initiative and are able

to exhibit goal-directed behavior; and

(4) social ability: co-operative or have the ability

to interact with other agents and objects with some

kind of language.

In addition to these, agents in the strong sense,

may possess additional characteristics that include:

(5) mobility: ability to move around;

(6) rationality: ability to perform in optimal

manner to achieve goals;

(7) benevolence: obey; and

(8) veracity: truthful.

4 ARCHITECTURE

We have employed an architecture where users (i.e.

consumers) specify requests and queries through

specified fashion ontology via user-interfaces. The

queries are routed to specialized agents for data

retrieval and their corresponding analysis of results.

Figure 2 depicts the overall architecture in terms of

its agents.

Consequently, we make use of this architecture

with three types of agent:

 UserAgent

 SearchAgent

 OntologyAgent

WEBIST 2005 - WEB INTERFACES AND APPLICATIONS

352

Each agent is in charge of different tasks such as

user interaction, ontology retrieval and metadata

search. Concretely, the architecture depicted in

Figure 2 includes three main agents which

communicate each other through information

messages according to FIPA-ACL standard (FIPA,

2002).

Figure 2: Architecture overview

Firstly, “UserAgent” constitutes the user's

intelligent gateway into this architecture. It uses

knowledge of the system's common domain model

(i.e. fashion ontology) to assist the consumer in

formulating queries and in displaying their results.

Furthermore, “UserAgent” interacts directly with the

agents “SearchAgent” and “OntologyAgent”.

Secondly, “SearchAgent” is responsible for

retrieving the metadata of documents. In this

architecture the RDF Data Query Language (RDQL)

(RDQL, 2004) was used to retrieve the metadata of

documents. We also used a search semantic tool like

JENA (Carroll et al., 2004) in order to search RDF

expressions.

Thirdly, the “OntologyAgent” provides an

overall knowledge of the e-fashion ontology so offer

answering queries about this domain and its

structure. The bulk of this e-commerce ontology is

devoted to define the common terms of fashion

clothing and organizing them as taxonomy.

In this stage of this research work we focus on

how to represent the selected ontology using mark-

up languages as described before. This will allow the

reusability and sharing of this knowledge between

different users (i.e. consumers). Thus, Figure 3

depicts fashion ontology (partial view) using RDF

graphs where some concepts are represented like

nodes, e.g. DesignSchool, DesignBranch, Style, etc.

Besides, links between nodes represent properties

like rdf:type and rdf:SubClassOf().

Figure 3: Fashion Ontology using RDF Graphs (partial

view)

On the other side, the subject ontology could be

expressed through RDF-S where several nodes are

represented in Figure 4.

Figure 4: Fashion Ontology using RDF-S (partial view)

5 CONCLUSIONS

This paper has introduced a few pointers about

impact of the growing intelligent agent technologies

and the Semantic Web on the phenomenon of e-

commerce. In particular, we have resumed how the

integration of ontologies and intelligent agents could

provide a new environment for e-commerce

applications. To conclude, the key advantages of our

outlined framework are:

(i) The use of techniques based on ontologies is

very important in order to enable concept

identification used more often in the knowledge

domain that e-commerce applications use. Then,

such techniques allow identifying core properties

and basic relationships existing among the concepts

in a given domain, becoming a powerful knowledge

representation metadata for knowledge repositories.

(ii) Well recognized technologies like RDF and

RDFS allow constructing refined knowledge

representations, even also providing a full set of

primitives in order to represent different kinds of

BUILDING E-COMMERCE WEB APPLICATIONS: Agent- and Ontology-based Interface Adaptivity

353

information resources which are part of any e-

commerce operation process.

(iii) Software agent technologies offer an

enormous potential of managing the information that

potential consumers use and generate. For that

reason, common tasks such as product search and

retrieval could be delegated to agents.

As a result, our prototype promises very good

results although our ongoing work clearly deals with

testing deeply our framework even also with more

powerful semantic languages like Ontology Web

Language (OWL) or the well-recognized Darpa

Agent Markup Language and Ontology Inference

Language (DAML-OIL). Another parallel future

goal being in progress is offering semantic

resolution and searching from new wireless

communications interface methods like PDA´s or

Screen Mobile Phones.

ACKNOWLEDGMENTS

This work has been partially funded by PBC-03-003

grant supported by Junta de Comunidades de

Castilla-La Mancha and by R.R.1256/04 grant

supported by University Miguel Hernández of Elche

– Bancaja.

REFERENCES

Bailin, S. & Truszkowski, W., 2001. Ontology negotiation

as a basis for opportunistic cooperation between

intelligent information agents. In Fifth International

Workshop on Cooperative Information Agents.

Springer. pp. 223-228.

Berners-Lee, T., Hendler, J., & Lassila, O., 2001. The

Semantic Web. Scientific American Journal pp. 33-43.

Brusilovsky, P. & Maybury, M., 2002. The Adaptive Web.

Communications of the ACM, vol. 45, no. 5, pp. 30-33.

Carroll, J. J., Dickinson, I., Dollin, C., Reynolds, D.,

Seaborne, A., & Wilkinson, K., 2004. Jena:

Implementing the Semantic Web Recommendations. In

13th World Wide Web Conference. ACM Press. pp. 74-

84.

FIPA, 2002. FIPA ACL Message Structure Specification.

Foundation for Intelligent Physical Agents. Available

online at http://www.fipa.org/specs/fipa00061/.

Hendler, J., 2001. Agents and the Semantic Web. IEEE

Intelligent Systems Journal, vol. 16, no. 2, pp. 30-37.

RDF, 1999. Resource Description Framework. Model and

Syntax Specification. W3C Recommendation.

Available online at http://www.w3.org/RDF/.

RDF-S, 2004. Resource Description Framework Schema.

Vocabulary Description Language. W3C

Recommendation. Available online at

http://www.w3.org/TR/rdf-schema/.

RDQL, 2004. A Query Language for RDF. W3C

recommendation. Available online at

http://www.w3.org/Submission/RDQL/.

SWCP, 2001. The Semantic Web Community Portal,

Markup Languages and Ontologies. Available online at

http://www.semanticweb.org/knowmarkup.html.

Tveit, A., 2001. A survey of Agent-Oriented Software

Engineering. In NTNU CSGSC Conference.

XML, 1998. The Extensible Markup Language. Available

online at http://www.w3.org/XML/.

WEBIST 2005 - WEB INTERFACES AND APPLICATIONS

354