Hontology: A Multilingual Ontology for the Accommodation Sector in

the Tourism Industry

Marcirio Silveira Chaves

and Larissa Freitas

and Renata Vieira

Business and Information Technology Research Center (BITREC), Universidade Atlˆantica, Lisbon, Portugal

Faculdade de Inform´atica, Pontif´ıcia Universidade Cat´olica do Rio Grande do Sul, Porto Alegre, Brazil

Keywords:

Knowledge Engineering, Ontology Development, Multilingual Ontology, Social Semantic Web, Accommo-

dation Sector.

Abstract:

Ontologies have been used to support both web agents reasoning and human decision making. However, on-

tology development is a new area and for some knowledge domains they are still rare. Although ontologies

had been developed in the context of Semantic Web, it is the Web 2.0 content that is actually pervasive on

the web. One of the properties of this content is to be multilingual, which requires multilingual resources to

deal with it. Online reviews are examples of multilingual texts provided by products and services consumers.

This paper presents a multilingual ontology for the accommodation sector. As a result, we deliver Hontol-

ogy, a freely available domain-speciﬁc ontology. Hontology reuses concepts of other vocabularies such as

Dbpedia.org and Schema.org. It is useful for a wide range of applications within the accommodation sector,

including ontology-based information extraction, text annotation and information visualisation.

1 INTRODUCTION

Multilingual Semantic Web is one of the most recent

challenges to applications dealing with web data, and

has gained more attention in the last years (Buitelaar

et al., 2010; Montiel-Ponsoda et al., 2011).

Ontologies have been adopted in order to put in

practice the vision of Semantic Web. Although the

knowledge engineering hard work and time consum-

ing tasks, the building of ontologies for a narrow do-

main of interest enables the development of semantic-

oriented applications. In the tourism industry, the ac-

commodation sector still lacks this kind of resource

to support agent reasoning. This sector contains sub-

domains such as hotel, hostels and apartments with

speciﬁc features which make the development of an

ontology much more complex.

Ontologies of the accommodations domain can be

used to support managerial decision making and end-

user applications. Instead of a few categories pro-

vided by Web 2.0 portals, e.g. TripAdvisor and Book-

ing.com, accommodation managers are more able to

ﬁnd speciﬁc information using an ontology. From the

user point of view, it is also easier to click on the con-

cept indoor swimming pool and read all the comments

mentioning it than to comb hundreds of comments to

ﬁnd opinions about indoor swimming pool. Unfortu-

nately, this last scenery probably is the current reality

in the most of Small and Medium Accommodations

in Europe.

The research presented in this paper has been

conducted of a framework to customer management.

This framework receives as input a set of online

reviews, uses Hontology to annotate them (Chaves

et al., 2012b) and presents this information in diverse

graphical formats to the decision making (Carvalho

and Chaves, 2012). To the best of our knowledge,

a robust, coherent and multilingual representation of

accommodationsector is lacking to bring into practice

the Semantic Web. The main purpose of this paper is

ﬁlling the gap of literature in describing the process

of building a multilingual ontology for the accommo-

dation sector. As a result, we deliver Hontology (H

stands for hotel, hostal and hostel), a freely available

domain-speciﬁc ontology in four languages: English,

Portuguese, Spanish and French.

The rest of this paper is structured as follows: Sec-

tion 2 describes the related works. Section 3 describes

the actors using Hontology and presents its require-

ments through Uniﬁed Modelling Language (UML)

use cases. Section 4 details the process of ontology

development and gives the statistics about Hontology.

Section 5 closes the paper with the ﬁnal remarks.

149

Silveira Chaves M., Freitas L. and Vieira R..

Hontology: A Multilingual Ontology for the Accommodation Sector in the Tourism Industry.

DOI: 10.5220/0004107401490154

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2012), pages 149-154

ISBN: 978-989-8565-30-3

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

2 RELATED WORK

Accommodation sector belongs to the tourism or

travel domain for which there are some ontologies de-

veloped, e.g. Mondeca (mondeca.com)and Harmonet

(harmonet.org). However, the vocabulary of available

tourism ontologies coversa limited set of concepts of-

ten describing the domain from different perspectives

due to the restricted application scope from which

the ontologies have been elicited (Barta et al., 2009).

Barta et al. (2009) produced the cDOTT ontology, a

common ontology for the tourism industry in order to

support the interoperability of the agents in low-level

operations.

Regarding multilingual ontologies, the Monnet

project (monnet-project.eu) - Multilingual Ontologies

for Networked Knowledge has presented some efforts

to put in pratice the vision of Multilingual Seman-

tic Web through of the production of ontologies to

the hydrographical domain (Aguado-De-Cea et al.,

2010). Aguado-De-Cea et al. (2010) use the label-

ing system supported by RDF(S) and OWL to in-

clude multilingual linguistic information in the ﬁrst

version of the ontology. Ou et al. (2008) devel-

oped QALL-ME, an ontology for the tourism domain,

which was linked to the multilingual terms in Eu-

roWordNet (http://www.illc.uva.nl/EuroWordNet/).

In Salim et al. (2010) an ontology for Islamic

Portal can retrieve information in three different lan-

guages (English, Malay, and Arabic). The framework

proposed to help users ﬁnd the information that they

need without worrying about the language, no matter

which language is used. The keywords of several lan-

guages are mapped to the same concept in an ontology

and are therefore given the same meaning.

Multilingual ontologies can also be used to struc-

ture an image database. Therefore, the problems re-

lated to the ambiguity could be solved. Popescu

(2007) developed a multilingual OWL ontology in-

cluding knowledge in English, Italian, and Spanish

which allow real-time processing when a user queries

for images.

Other example of application involving multilin-

gual ontologies is Jung (2011), who proposes a multi-

agent system for building indirect alignment between

multilingual ontologies. This system was tested with

multilingual ontologies written in English, Korean,

and Swedish languages. A case study was realized

in tourism business domain.

3 HONTOLOGY

REQUIREMENTS

Hontology was initially created considering a scene-

rio of applications processing reviews about accom-

modations on Web sites 2.0 and decision support sys-

tems (DSS). The conception of an ontology should

also take into account the users point of view, which in

case of Hontology includes at least six types of users:

Database Administrator: is interested in Hontol-

ogy in order to better deﬁne the database schema

to store instances or individuals of Hontology

(e.g. reviews, features, polarity and intensity of

the reviews). In Description Logic (Baader et al.,

2007), the set of these instances is denominated

A-Box, while the schema is called T-Box.

Application Developer: is the person who uses the

content of Hontology to develop applications that

support decision making.

Prospect: is a user seeking for accommodations.

She is interested in common features of an ac-

commodation such as cleaning and staff. These

features allow her to make comparisons among

accommodations based on her speciﬁc needs.

Guest: is a former or current customer of an accom-

modation. Further the same interest of a prospect,

a guest can be interested in providing feedback.

Domain Expert: is the person responsible to main-

tain the concepts and relationships in the ontol-

ogy. Her tasks include the matching of concepts

to existing ontologies and augmenting Hontology

according to the new requirements.

Accommodation Manager: includes managers in

strategic and managerial levels of the organization

such as the Owner of the Hotel, Marketing man-

agers and Service managers.

4 KNOWLEDGE ENGINEERING

Hontology was developed in seven phases as de-

scribed in Chaves and Trojahn (2010).

1. Identify existing ontologies on related domains;

2. Select the main concepts and properties;

3. Organize concepts and properties hierarchically

into categories;

4. Manually translation of the ontology introducing

the labels in Portuguese, Spanish and French;

5. Expand concepts and properties based on online

reviews manually evaluated;

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

150

6. Translate the new concepts and properties;

7. Export the ontology in several formats.

Domain experts updated Hontology, which in-

cludes the mapping of the concepts to existing ontolo-

gies. Classiﬁcations of accommodations vary from

country to country and in some continents it is possi-

ble to ﬁnd speciﬁc types of accommodations. One of

the guidelines followed in the development of Hon-

tology was that proposed in EC (2009). Chaves et al.

(2012b) describe an analysis of the reviews about

Small and Medium Hotels in Portugal. This analysis

was used as input to build the extension of Hontology,

which is detailed in the next sections.

4.1 Main Concepts

The deﬁnition of the top-level concepts was based on

the needs of the main actor. From these concepts, the

actors have a broad and depth vision of the domain,

which they are looking for information.

Accommodation. This is the substitute concept of

Hospitality which was the concept used in the ﬁrst

version of Hontology. Instead of extend Hontol-

ogy to cover restaurants, transportation and event

planning, we extend the set of subclasses to spec-

ify the different types of Accommodation. The

design of the concept Accommodation is based on

the categories provided by EC (2009). Actually,

Accommodation contains sixteen subconcepts in-

cluding six kinds of hotel under the concept Ho-

tel: Bunker, Capsule, Cave, Ice, Tree House, and

Under Water. This list can be extended depend-

ing on the applications needs. All the sixteen sub-

concepts of Accommodation are listed in the ﬁrst

column of Table 1.

Facility. This concept was split in six subconcepts:

External Facility, Internal Facility, Room Facil-

ity, Bathroom Facility, Driver ans Wheel Chair

Accessible. The last two ones could be into In-

ternal Facility, but we decide to create a separate

subconcept in order to facilitate the navigation. It

is worth remembering that Internal Facility con-

tains thirty subconcepts.

Room. Contains subconcepts for the main accom-

modation types, that is, Apartment Room, Hos-

tel Room and Hotel Room as depicted in Figure

1. Hotel Room contains nine subconcepts includ-

ing Double Room, Suite Room and Family Room.

Family Room is still subdivided into three con-

cepts: Family Junior Suite with Sea View, Fam-

ily Room with Balcony and Family Suite with Sea

View. Some special cases in the knowledge en-

gineering evolve typical properties that are repre-

Figure 1: Representation of Room types in Hontology.

Figure 2: Representation of Service and Staff in Hontology.

sented as concepts in Hontology, e.g. with Bal-

cony and with Sea View.

Service/Staff. Figure 2 presents the subconcepts of

the concepts Service and Staff. Most of the service

offered by an accommodation owns a responsible

to performit that is why we decide to present these

concepts together.

Guest Type. Figure 3 depicts the main kinds of

Guest Type. These types are commonly used in

travel 2.0 sites in order to classify guests accord-

ing to a pre-deﬁned proﬁle.

Other concepts to mention include Design, Meal,

Points of Interest, Price, Rating and Staff.Price usu-

ally is modelled as a property in ontologies, but some

of the uses of Hontology may require that price must

be explicit as a concept given its relevance to the ap-

plication. For instance, hotel managers can be in-

terested in reviews mentioning prices of the speciﬁc

services, such as BarPrice, BreakfastPrice, CoffeeP-

rice, CotPrice, InternetPrice, ParkingPrice, Restau-

Hontology:AMultilingualOntologyfortheAccommodationSectorintheTourismIndustry

151

Figure 3: Representation of Guest Type in Hontology.

rantPrice, and RoomPrice. The same principle was

applied to the feature Timetable, which is represented

as a conceptwith four subconcepts: Indoor Swimming

Pool Timetable, Outdoor Swimming Pool Timetable,

Restaurant Timetable, and Spa Centre Timetable.

4.2 Exploring Ontology Constructs

This section describes the main constructs used to

represent the multilingual and logical aspects in the

accommodation sector.

rdfs:label, skos:prefLabel. Whenever Hontology is

exported to RDF(S) or SKOS, multiligual infor-

mation is codiﬁed using the constructs rdfs:label

and skos:prefLabel, respectively. Into these con-

structs, we use the attribute xml:lang with the

respective language as a value. The construct

skos:prefLabel stores the preferred term in a lan-

guage.

owl:equivalentClass, owl:sameAs, rdfs:subClassOf.

These labels, at conceptual level, can help the

conceptual cross-lingual mappings. That is,

concepts from different ontologies described in

different languages can be semantically related

by using ontology constructs, either to represent

taxonomical relations, e.g. owl:equivalentClass,

owl:sameAs, rdfs:subClassOf (Gracia et al.,

2011). In the version of Hontology described in

this paper, we use only the owl:equivalentClass,

owl:sameAs constructs to map concepts to other

ontologies (see details in Section 4.4). We do not

deal with hierarchical mappings here.

owl:disjointWith. Hontology also handles com-

plex classes througth the owl:disjointWith con-

struct. For example, all concepts under the

concepts Accommodation, GuestType, Room,

Meal, PointsOfInterest, Price, Service, Staff and

Timetable are codiﬁed as disjoint.

4.3 Multilingual Aspects

In addition to provide a new ontology for the accom-

modation sector, we also make an effort to handle

multilingual concepts in order to foster the Multilin-

gual Semantic Web. We start from four languages, i.e.

English, Portuguese, Spanish and French. In addition

to the main concepts in each language about the ac-

commodation domain, we also store variants when-

ever they exists. For Portuguese, for instance, we

store the variants of European and Brazilian terms.

Examples include the concept Playground which is

called Prac¸a de Brinquedos in Brazil and Parque In-

fantil in the European Portuguese, and Address which

is called Enderec¸o in Brazil and Morada in Portugal.

Although, XML does not support the explicit

markup of variants, applications using Hontology

can ﬁnd this information as a value of the attribute

xml:lang. This information separately facilitate the

storage in linguistic repositories, such as that pro-

posed in Montiel-Ponsoda et al. (2008). We intend

to extend the use of variants in other languages than

Portuguese, according to the requirements of the ap-

plications using Hontology.

4.4 Knowledge Reuse

One of the principles followed along the development

of Hontology was the reuse of the existing resources.

Although ontologies in the speciﬁc domain of accom-

modations are rare or even inexistent, we can ﬁnd

concepts of this domain formalised in broader do-

mains such as tourism.

The several actors interested in the accommoda-

tion information share common concepts that need

to be represented or formalised in a single structure

or aligned with existing vocabularies. We consid-

ered concepts from other public ontologies QALL-

ME (Ou et al., 2008), Schema.org and Dbpedia.org.

We then aligned our ontology concepts with these

other public resources according to Table 1. More-

over, the subconcepts of the concept Accommodation

in Hontology also include: Apartment, Botel (Boat

Hotel), BungalowAndCamping, GuestHouse, Holi-

daySettlement, Hostal, Inn, Pension, and OtherAc-

commodation.

Table 2 presents the mappgins between the con-

cepts of Hontology and QALL-ME. Figure 4 depicts

the representation of the concepts Facility and Room-

Facility in both Hontology and QALL-ME ontolo-

gies. Hontology maps the concept Facility to the ex-

isting one in QALL-ME and extends its subconcepts.

BathroomFacility owns 10 subconcepts, ExternalFa-

cility contains nine subconcepts, InternalFacility has

29 subconcepts and RoomFacility owns 22 subcon-

cepts. It is worth mentioning that QALL-ME was

designed to cover the wide domain of tourism and

Hontology was built to the speciﬁc sector of accom-

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

152

Table 1: Alignment between the subconcepts of Accommodation in Hontology and QALL-ME, Lodging Business in

Schema.org and Tourist Accommodations in Dbpedia.org.

Hontology QALL-ME Schema.org Dbpedia.org

BedAndBreakfast BedAndBreakfast BedAndBreakfast BedAndBreakfast

Hostel Hostel Hostel Hostel

Hotel Hotel Hotel Hotel

Motel Motel Motel

Chalet Chalet

Cottage Cottage

Resort Resort Resort

Table 2: Alignment between the concepts of Hontology and

QALL-ME.

Hontology QALL-ME

Accommodation Accommodation

Airport Airport

BusStation BusStation

Country Country

Facility Facility

Location Location

MetroStation MetroStation

Price Price

Restaurant Restaurant

RestaurantPrice GastroPrice

RoomPrice RoomPrice

Stadium Stadium

Theatre Theatre

TrainStation TrainStation

modations. For this reason, Hontology speciﬁes the

subconcepts of Facility in more detail.

Moreover, it is important to note that the concept

RoomFacility occurs in both ontologies but it is not

aligned. The subconcepts in QALL-ME represent

types of room facility while the subconcepts in Hon-

tology represent the infrastructure that is present or in

lack in a room.

4.5 Hontology Metrics

Hontology contains 282 concepts categorized into 16

top-levelconcepts. The concept hierarchy has a maxi-

mum depth of 5. Table 3 presents some metrics about

Hontology. It is worth mentioning that all concepts

and properties are deﬁned in English, Portuguese,

Spanish and French. Along the development of this

ontology, we observe that Hontology contains the fol-

lowing qualities: i) indeﬁnite expandability, since it

remains consistent with increasing content. ii) content

and context independence, since any kind of concept

can ﬁnd its place and iii) specify different levels of

granularity (levels range from 1 to 5 – see an example

in Figure 1).

Figure 4: Representation of the concept Facility in Hontol-

ogy (left) and in QALL-ME (right).

5 FINAL REMARKS

This paper introduces a new resource to put in

practice the Multilingual Semantic Web. Hontol-

ogy, a robust multilingual ontology for the ac-

commodation sector, helps to pave the path for

reaching this Web with meaning. Considering

that Hontology can be useful for a wide range of

knowledge-based systems, we make it freely available

from http://ontolp.inf.pucrs.br/Recursos/downloads-

Hontology.php. Preliminary experiments using Hon-

tology were performed for Portuguese in Chaves et al.

(2012a). In this sense, we provide a contribution

knowledge enginnering, ontology development and

domain applications. Future works include the task

of ontology population, where we plan to implement

algorithms to recognise concepts of the Hontology in

the reviews and make them the appropriate instances

to the concepts.

Hontology:AMultilingualOntologyfortheAccommodationSectorintheTourismIndustry

153

Table 3: Hontology Metrics.

Metrics Values

Number of Concepts 282

Number of Object Properties 8

Number of Data Properties 31

Concept Axioms

Subconcept axioms 272

Equivalent concepts axioms 4

Disjoint concepts axioms 93

Object Property Axioms

Functional object property axioms 5

Object property domain axioms 9

Object property range axioms 9

Data Property Axioms

Functional data property axioms 12

Object data domain axioms 17

Object data range axioms 1

ACKNOWLEDGEMENTS

We thank Elizabeth Carvalho and C´assia Trojahn for

helping in the Spanish and French ontology transla-

tion, respectively. We thank the Brazilian funding

agency CAPES for the scholarship granted.

REFERENCES

Aguado-De-Cea, G., G´omez-P´erez, A., Montiel-Ponsoda,

E., and Vilches-Bl´azquez, L. M. (2010). Rivi`ere or

ﬂeuve? modelling multilinguality in the hydrograph-

ical domain. In MultiLingual Semantic Web (MSW

2010).

Baader, F., Calvanese, D., McGuinness, D. L., Nardi, D.,

and Patel-Schneider, P. F., editors (2007). The De-

scription Logic Handbook: Theory, Implementation

and Applications. Cambridge University Press, 2nd

edition.

Barta, R., Feilmayr, C., Pr¨oll, B., Gr¨un, C., and Werthner,

H. (2009). Covering the semantic space of tourism: an

approach based on modularized ontologies. In Proc.

of the 1st Workshop on Context, Information and On-

tologies, pages 1:1–1:8, New York, NY, USA. ACM.

Buitelaar, P., Cimiano, P., and Montiel-Ponsoda, E., editors

(2010). 1st International Workshop on the Multilin-

gual Semantic Web. CEUR-WS.org.

Carvalho, E. and Chaves, M. S. (2012). Exploring user-

generated data visualization in the accommodation

sector. In 16th International Conference Information

Visualisation, Montpellier, France, 10-13 July. IEEE.

Chaves, M. S., Freitas, L., Souza, M., and Vieira, R.

(2012a). PIRPO: An algorithm to deal with polarity

in Portuguese online reviews from the accommodation

sector. In 17th International conference on Applica-

tions of Natural Language Processing to Information

Systems (NLDB), Groningen, The Netherlands, 26-28

June.

Chaves, M. S., Gomes, R., and Pedron, C. (2012b). Deci-

sion making based on web 2.0 data: The small and

medium hotel management. In Proceedings of the

20th European Conference on Information Systems

(ECIS), Barcelona, Spain, 10-13 June.

Chaves, M. S. and Trojahn, C. (2010). Towards a multilin-

gual ontology for ontology-driven content mining in

social web sites. In Proceedings of the ISWC 2010

Workshops, Volume I, 1st International Workshop on

Cross-Cultural and Cross-Lingual Aspects of the Se-

mantic Web., Shanghai, China.

EC (2009). Classiﬁcation of Hotel Establishment within the

EU. Available at http://ec.europa.eu/consumers/ecc/

docs/hotel establishment classiﬁcation EU en.pdf.

Gracia, J., Montiel-Ponsoda, E., Cimiano, P., G´omez-P´erez,

A., Buitelaar, P., and McCrae, J. (2011). Challenges

for the multilingual web of data. Journal of Web Se-

mantics: Science, Services and Agents on the WWW,

11(0).

Jung, J. J. (2011). Exploiting multi-agent platform for in-

direct alignment between multilingual ontologies: A

case study on tourism business. Expert Systems with

Applications, 38(5):5774–5780.

Montiel-Ponsoda, E., Aguado de Cea, G., G´omez-P´erez,

A., and Peters, W. (2008). Modelling Multilinguality

in Ontologies. In Coling 2008: Companion volume:

Posters, pages 67–70, Manchester, UK.

Montiel-Ponsoda, E., McCrae, J., Buitelaar, P., and Cimi-

ano, P., editors (2011). Proceedings of the 2nd Inter-

national Workshop on the Multilingual Semantic Web.

Ou, S., Pekar, V., Orasan, C., Spurk, C., and Negri, M.

(2008). Development and alignment of a domain-

speciﬁc ontology for question answering. In Proc. of

the LREC, 26 May - 1 June, Marrakech, Morocco.

Popescu, A. (2007). Image retrieval using a multilingual

ontology. In Computer-Assisted Information Retrieval

(Recherche d’Information et ses Applications) - RIAO

2007, 8th International Conference, Carnegie Mellon

University, Pittsburgh, PA, USA, May 30 - June 1,

2007. Proceedings, CD-ROM.

Salim, J., Hashim, S., and Aris, A. (2010). A framework

for building multilingual ontologies for Islamic por-

tal. In Proc. of the International Symposium in In-

formation Technology (ITSim 2010), volume 3, pages

1302–1307.

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

154