445 Semantic Web and E-Tourism/445 Danica Damljanović* Department of Computer Science, University of Sheffield Regent Court 211 Portobello Street S1 4DP Sheffield United Kingdom [email protected] +44-114-222-1800 +44-114-222-1810 Vladan Devedžić FON - School of Business Administration, University of Belgrade POB 52 Jove Ilića 154 11000 Belgrade Serbia [email protected] +381-11-3950-853 +381-11-461221

Semantic Web and E-Tourism INTRODUCTION Offering tourist services on the Internet has become a great business over the past few years. Heung (2003) revealed that approximately 30% of travelers use the Internet for reservation or purchase of travel products or services. Classic sites of tourist agencies enable users to view and search for certain destinations and book and pay for vacation packages. At a higher level of sophistication are tourism Web portals, which integrate the offers of many tourist agencies and enable searching from one point on the Web. Still, when using this kind of systems one is forced to spend a lot of time analyzing Web content with destinations that match his/her wishes. This problem is identified by Hepp, Siorpaes and Bachlechner (2006) as the “needle in the haystack” problem. Applying artificial intelligence (AI) techniques in E-Tourism could help resolve this problem by providing: 1. data that are semantically enriched, structured, and thus represented in a machine readable form; 2. easy integration of tourist sources from different applications; 3. personalization of sites: the content can be created according to the user profile; 4. improved system interactivity. As an example of using AI in E-Tourism, we present Travel Guides – a prototype system that offers tourists complete information about numerous destinations. They can search destinations by using several criteria (e.g., accommodation type, food service, budget, activities during vacation, and user interests: sports, shopping, clubbing, art, museum, monuments, etc.). He/She can also read about the weather forecast and events in the destination.

2

In a way, Travel Guides complements traditional information systems of tourist agencies. These systems require a lot of maintenance effort in order to keep the huge amount of data about tourist destinations up-to-date. Travel Guides is created to minimize the user's input and his/her need to filter information. It shows how usage of semantically enriched data in a machine readable form can •

increase interoperability in the area of tourism,

•

decrease maintenance efforts of tourist agents, and

•

offer tourists a better service.

Nowadays there are just a few E-Tourism systems that use AI techniques. We briefly discuss them in the next section. In this article we explain why it would be good to use such techniques and how Travel Guides does it. Specifically, using Semantic Web technologies in the area of tourism can improve already existing systems (which are mostly available online) that do not use Semantic Web techniques yet. Likewise, the Semantic Web approach can help decrease the maintenance efforts required for existing E-Tourism systems and ease the process of searching for vacation packages. Travel Guides was initially developed as a large-scale expert system. Over time, it has evolved into a modern Semantic Web application. BACKGROUND According to Aichholzer, Spitzenberger & Winkler (2003), E-Tourism comprises electronic services which include: •

information services (e.g. destination, hotel information);

•

communication services (e.g. discussion forum);

•

transaction services (e.g. booking).

3

Transaction services are offered at many places on the Web , such as Expedia, Travelocity, etc. These Websites include some of the information services, but for complete details about certain destination (e.g., activities, climate, monuments, and events) one must search for other sources. Some Websites even help in planning the whole itinerary (e.g. HomeAndAbroad). Apparently, there is an “information gap” between these online services, and no interoperability. Semantic Web technologies can be used to overcome this problem and thus increase the quality of ETourism. Cunningham (2002) presents GATE (General Architecture for Text Engineering) as an infrastructure for developing and deploying software components that process human language. It can annotate documents and recognize concepts such as: locations, persons, organizations and dates. GATE can annotate documents with respect to a particular ontology. Some of the recently built C, like KIM (Popov, Kiryakov, Ognyanoff, Manov & Kirilov, 2004), use GATE for information extraction and retrieval. Similar to other AI technologies, Semantic Web is not frequently used in real-time tourism applications. Integrating AI tools into mainstream applications can result in benefits to both sides (Djuric, Devedzic & Gasevic, 2007). Standard organizations like the Internet Engineering Task Force and the World Wide Web Consortium (W3C), are making major efforts at developing languages for sharing meaning (Shadbolt, Berners-Lee & Hall, 2006). Speaking at the WWW2006 conference in Edinburgh in May 2006, W3C director Tim Berners-Lee, pointed out that Semantic Web has all the standards and technologies it needs to succeed and that it was time for Web developers and content producers to start using semantic languages in addition to HTML (Bennett, 2006). Cardoso (2006a) addresses the lack of standards in the tourism domain: the prices for tourism

4

activities are expressed in different currencies; the time units also do not follow the standards. He argues that use of Semantic Web and ontologies could overcome this problem. In (Cardoso, 2006b) he describes the ontology developed to achieve integration and interoperability through the use of a shared vocabulary and meanings for terms with respect to other terms in the area of tourism. His system creates vacation packages dynamically using previously annotated data in respect to the ontology. This is performed with a service that builds itinerary by combining user preferences with flights, car rentals, hotel, and activities in a single price. Similar to this Jakkilinki, Georgievski & Sharda (2007) presents a tool for tour planning that is intelligent in the meaning that it generates travel plans by matching user preferences and available tourist offers from different travel agents in respect to the ontology which enables reasoning. To use Semantic Web in E-Tourism two approaches could be applied. One is to make applications from scratch, based on the existing standards. The other one is to enrich already existing content with annotations based on ontology. The first approach is not cost-effective for tourist agencies. Although the second approach sounds more reasonable, it seems that there are not enough data in the domain of tourism on the Web. Hepp et al. (2006) made a research in this field using a sample of 100 accommodations in Austria. Their results showed that neither some of the hotels had their Web pages, nor the biggest Austrian portal for E-Tourism (Tiscover) had any information about them. An additional problem they noticed was the incompleteness of the details such as the availability of the accommodation and the prices. Many E-Tourism portals store their data internally, and not on the Web. This means that even a perfect annotation of the Web content wouldn't be sufficient enough hence it is limited to persistently published information (Hepp, 2006). To exceed this problem Stojanovic, Stojanovic & Volz (2002) developed a mapping mechanism for migrating relational database schemas into

5

ontologies in order to form the conceptual backbone for metadata annotations which are automatically created from the database. A better approach would be to use Semantic Web services, e.g. Web Service Modeling Ontology - WSMO (Roman et al., 2005) or OWL-based Web service ontology - OWL-S (Smith & Alesso, 2005). Dell’erba, Fodor, Hopken, & Werthner (2005) present the Harmonize project that integrates Semantic Web technologies and merge tourist electronic markets using ontology as a mediator. Their ontology was taken over by the E-Tourism Working Group (2004) at Digital Enterprise Research Institute (DERI). This group plans to develop an advanced E-Tourism Semantic Web portal, which will connect the customers and virtual travel agents. This portal could be of importance to the travel industry in Austria, whereas for the rest of the world it could be an example of using Semantic Web technologies in a real business system. In 2001, the industry tried to address the interoperability issue by forming a consortium called the Open Travel Alliance. OTA is producing XML specifications (schemas) for messages to be exchanged between the trading partners, e.g. availability checking, booking, rental, reservation, query services, insurance, etc. The precondition for this improvement method to succeed is that each travel agent’s application can produce and consume OTA-compliant messages. Dogac et al. (2004) present the SATINE project as a peer-to-peer network that enables peers to deploy their semantically-enriched travel Web services and allows others to discover these services semantically. In addition to attempts to semantically enrich tourism sources on the Internet, it has been very popular to develop a Location Based Tourism Systems (LBTS). LBTS are computerized systems that depend on an automated location of a target which either deliver or collect information. Currently LBTS applications are being used by mobile phones, iPods, PDAs (Hawking et al.,

6

2005). LBTS provide search for hotels and ATM machines near by the user’s current location and additional information when the user visits a city for the first time. An example of such a system is “Mobility Agent” (Edwards, Blythe, Scott & Weihong-Guo, 2006). This system delivers Internet-based travel and tourism-related services through fixed and mobile devices. Intelligent agent technology (Devedzic, 2003) was used to provide European visitors the dynamic, mobile, personalized, location-based information and services, especially related to travel in complex urban environments. Kanellopoulos & Kotsiantis (2006, p. 86) predict that “in a collaborative travel environment, agents will be essential in addressing the issues of security, negotiation, personalisation and Web Service procurement”. TRAVEL GUIDES Travel Guides is a Semantic Web portal in the area of tourism. Designed with the idea to gather all vacation packages from different tourist agencies, this system is built to help searching for a 'perfect' vacation package. It is also personalized, so that its content is adapted to the user regarding his/her interests and activities. It is an example of using Semantic Web to increase the interoperability, provide better interaction, and intelligent reasoning in the domain of tourism. Requirements for an E-Tourism intelligent portal To make a tourism portal capable of intelligent reasoning, it is necessary to: •

build some initial knowledge in the system,

•

introduce user profiles, and

•

maintain the knowledge automatically during the user's interaction with the system.

That way, the knowledge collected in a machine-readable form reduces the user's input and improves search.

7

In addition, the portal should be able to gather useful Web content and extract information of potential interest to the user. In order to produce knowledge in a machine-readable form, it is necessary to develop and use a set of ontologies to represent all important concepts and their relations. Ontologies also enable knowledge sharing and reuse between applications, and development of intelligent Web services by using Semantic Web technologies and tools. In case of search-intensive applications like tourism portals, ontologies bring up an important feature of semantic search. Unlike keyword-based syntactic search, ontology-supported semantic search returns better-quality information, because of the possibility to find the pages that contain semantically similar albeit possibly syntactically different text. A related concept is that of semantic interoperability. It is best understood by contrasting it to syntactic interoperability: syntactic interoperability is about parsing the data, while semantic interoperability means using ontologies to define mappings between data and known concepts. For example, a tourist agency may use the term "day trip" on its Website, whereas another one may use "1-day excursion". Ontology from the domain of tourism may be used to make the equivalence mapping between the two terms, thus enabling the portal to treat the related pieces of information equally. User profiles Personalization implies adapting content to a particular user and enhances the interactivity of the system. User profiles are based on the data that the system has about the user. These data are provided in two ways: 1. The user is willing to fill the forms where he/she inputs data about him/herself: gender, birth date, social data, address, profession, education, languages, interests and activities, budget, visited destinations (cities, countries).

8

2. The system collects data about the user’s interests and preferences while the user is visiting the portal and reading about or searching for vacation packages. Each time the user clicks on some of the vacation package details, the system stores his/her action in a database. With the user profile, the system can adapt its look-and-feel and, more importantly, its content to the user. For example, the results of search for a destination (based on the requested criteria) can be additionally filtered according to the user profile, Figure 1. Each destination is identified by its latitude and longitude. When the system is aware of the user's location, the search results could be sorted by destinations that are physically closer to the user.

Figure 1: An adaptive portal using the user's profile Whenever possible, the user profile extension and maintenance is performed dynamically by the system itself without the user's explicit intervention, using specific heuristics. There is a need to "observe" the user constantly and to use a built-in reasoner to infer the user's preferences and intentions from the observations, i.e. to create and maintain a valid user profile based on his/her behavior.

9

The portal architecture Figure 2 depicts the architecture of the Travel Guides portal for tourism management. The User interface component collects data from the user and sends them to the Controller. Controller manages the requests from the interface and fires appropriate actions. The User Manager takes care of the user data. It uses the User DAO (Data Access Object) to store or fetch the needed data from/to a database. These data are user details that are not subject to frequent change and are not important for determining the user profile: the username, password, first name, last name, address, birth date, phone and email. The User Profile Expert is aware of the User ontology and also of the User profile knowledge base that contains instances of classes and relations from the User ontology. The Travel Manager is responsible for fetching, storing and updating the data related to vacation packages. Storing and retrieving data from the database is performed by the Vacation Package DAO. The data stored in the database are those that are subject to frequent changes and are not important in the process of reasoning: start date, end date, prices (accommodation price, food service price, transport price), benefits, discounts, and documents that contain textual descriptions with details about the vacation packages. Some of these data are used in search queries as constraints. Actually, retrieving a 'perfect' vacation package is performed in two steps: 1.

Matching the user's wishes to certain destinations – the user profile is matched with

certain types of destinations. This is performed by the Travel Offer Expert and the World Expert. 2.

The list of destinations retrieved in the first step is filtered using the constraints the user

may want to set (for example, the start/end dates of the vacation). The filtering is done by the Vacation Package DAO. The Travel Manager aggregates the Tourist Offer Expert and World Expert components, which

10

include inference engines. These inference engines are aware of the Travel and World ontology respectively. Their main role is to validate the content in the knowledge bases. The System Scheduler is running daily at scheduled times. It takes data about the type of destinations that the user has visited and updates his/her user profile. Those data are fetched from the logged user's activities in the relational database, semantically processed and stored in a temporal knowledge base. After the temporal knowledge base is validated, it is merged with an already existing user-profile knowledge base. This validation is performed by an inference engine built inside the User Profile Expert. A similar process is performed to update tourist offers in the knowledge base. This happens when a travel agent updates data about new vacation packages, as well as when the portal administrator updates the World knowledge base (see Figure 2).

11

Figure 2: The architecture of the Travel Guides portal for tourism management The relations and restrictions stored in Travel Guides ontologies represent the core of this system. Ontologies allow machine-supported travel-related data interpretation and integration (Kanellopoulos, Kotsiantis & Pintelas, 2006). The World ontology contains concepts and relations from the real world: geographical terms, locations with coordinates, land types, time and date, time zone, currency, languages, and all other terms that are expressing some concepts that are in a way related to tourism or tourists, but not to vacation packages that could be offered by tourist agencies. Since Travel Guides is intended to support semantic annotation, indexing, and retrieval of documents, this ontology is also meant to contain the general concepts necessary for expressing semantic annotation, indexing, and retrieval.

12

The User ontology is meant to contain data about the users – the travelers who visit the Travel Guides portal. This ontology describes user interests and activities, age groups, favorite travel companies, and other data related to user profiles. Each user can have one or more user profiles, depending on his/her behavior while visiting the portal, and also depending on the data he/she has provided to the system. The Travel (Tourism) ontology includes all terms being specific to types of vacations, traveler types, and vacation packages offered in tourist agencies and being important to travelers, like the type of accommodation, food service type, transport service, type of room in a hotel, and the like. It is an ontology that makes an indirect connection between users and destinations. The ontologies are implemented in OWL (Antoniou & Harmelen, 2004) and developed using the Protégé tool (Horridge, Knublauch, Rector, Stevens & Wroe, 2004). Travel Guides is designed so that all users contribute to the creation and updating of the knowledge base. Each group is contributing to the knowledge base in its own way: •

End users (tourists) feed the system with their personal data, which then get analyzed by

the system in order to create/update user profiles. The system also uses logged data about each user's activities (mouse clicks) when updating the user's profile. •

Tourist agents are creating vacation packages and similar offers in tourist agencies. They

feed and update the knowledge base with new information about destinations, arrangements, excursions, etc. To do this, they fill a form about a destination which includes fields like name, accommodation, hotel name, parking provided by the hotel, swimming pool inside the hotel, activities, transport service, etc. •

Portal administrators – they mediate the knowledge base updates with destinations not

covered by the tourist agencies. The idea is that tourist agents can then use those updates as the

13

basis for creating new vacation packages and other tourist offers. To alleviate the creation, extensions, and maintenance of this part of the knowledge base, Travel Guides exploits WordNet1 (Fellbaum, 1998), an open-source knowledge base that describes a number of concepts and terms. Some of these were copied to the Travel Guides knowledge base to avoid manual entrance of static (permanent) data about various destinations. A special tool is developed for copying instances of WordNet classes (concepts) and relations to instances of classes and relations of the Travel Guides World ontology. FUTURE TRENDS There is room for future improvements in several directions: •

The use of the Semantic Web services.

•

The system is designed to possibly include information about all places in the world and

all vacation packages. To achieve this, it is necessary to feed it and save a great amount of data. The system's prototype described here includes a limited collection of vacation packages, which must be extended and updated. •

Simplification of the document annotation process. Currently, it requires the knowledge

and understanding of the GATE. •

Finally, Travel Guides supports only management of hotel accommodation. This should

be expanded with hostels, campgrounds, and private apartments. CONCLUSION Artificial intelligence can play a significant role in improving E-Tourism portals. The use of Semantic Web technologies can increase interoperability in the area of tourism although the agreement of all involved parties about using the standards is essential. Without this agreement, 1

Wordnet 2.1: http://wordnet.princeton.edu/

14

the only useful thing that could be achieved with annotation process would be to analyze sites with content about countries, cities, beaches, etc. Most of the data regarding vacation packages are internally stored in the tourism portals, hence only the usage of intelligent Web services would help their retrieval and representation in a machine readable form. Standardization of the way all tourist services are representing the data would speed up the process of their integration. This would ease searching for tourist deals from one place. The integration of geographical data would also decrease efforts of tourist agents who are responsible to feed the system and keep data up-to-date. If these data would be centralized in a repository available to tourist agents, this would significantly decrease the maintenance efforts. The Travel Guides system is built to solve the problem of distributed tourist sources and to help users find a 'perfect' vacation package quickly. The system includes intelligent components that perform reasoning and have some built-in heuristics. It is based on a number of ontologies, which support the process of adding semantics to data. The data are semantically enriched, which means that they can be understood by machines. The system is also personalized in order to adapt its content to each user. The more a user visits the portal, the more the system “knows” about him/her. As this system is built using the latest Java technologies for building Web applications widely used in existing online tourism information systems nowadays, this prototype system is to show that such mainstream systems can benefit from integrating the semantic web components. REFERENCES Aichholzer, G., Spitzenberger, M. & Winkler, R. (2003, April). eTourism Strategic Guideline 6, PRISMA – Providing Innovative Service Models and Assessment, April, Vienna: Institute of

15

Technology Assessment, Austrian Academy of Sciences. [Online]. Retrieved January 13, 2007 from: http://www.prisma-eu.net/deliverables/sg6tourism.pdf Antoniou, G., Harmelen, F. V. (2004). Web Ontology Language: OWL. In Staab, S.,Studer, R. (Eds.): Handbook on Ontologies (International Handbooks on Information Systems) (pp. 67-92), Springer. Bennett, J. (2006, May 25). The Semantic Web is upon us, says Berners-Lee. Silicon.com research panel: WebWatch. [Online]. Retrieved January 3, 2007 from: http://networks.silicon.com/webwatch/0,39024876,39159122,00.htm Cardoso, J. (2006a). Developing Dynamic Packaging Systems using Semantic Web Technologies. Transactions on Information Science and Applications. 3(4). 729-736. Cardoso, J. (2006b). Developing An Owl Ontology For e-Tourism. In Cardoso, J. & Sheth, P. A. (Eds.). Semantic Web Services, Processes and Applications (pp. 247-282), Springer. Cunningham, H. (2002). GATE, a General Architecture for Text Engineering. Computers and the Humanities. 36 (2). 223–254. Dell’erba, M., Fodor, O., Hopken, W. & Werthner, H. (2005). Exploiting Semantic Web technologies for harmonizing E-Markets. Information Technology & Tourism. 7(3-4). 201219(19). Devedzic, V. (2003). Intelligent Information Systems. Digit, FON, Beograd (in Serbian). Djuric, D., Devedzic, V. & Gasevic, D. (2007). Adopting Software Engineering Trends in AI. IEEE Intelligent Systems. 22(1). 59-66.

16

Dogac, A., Kabak ,Y., Laleci, G., Sinir, S., Yildiz, A. & Tumer, A. (2004). SATINE Project : Exploiting Web Services in the Travel Industry. eChallenges 2004 (e-2004), 27 - 29 October 2004, Vienna, Austria. Edwards, S. J., Blythe, P. T., Scott, S. & Weihong-Guo, A. (2006). Tourist Information Delivered Through Mobile Devices: Findings from the Image. Information Technology & Tourism. 8 (1). 31-46(16). E-Tourism Working Group (2004). Ontology Collection in view of an E-Tourism Portal. October, 2004. [Online]. Retrieved January 13, 2007 from: http://138.232.65.141/deri_at/research/projects/e-tourism/2004/d10/v0.2/20041005/ Fellbaum, C. (1998). WordNet - An Electronic Lexical Database. The MIT Press. Hawking, P., Stein, A., Zeleznikow, J., Pramod, S., Devon, N., Dawson, L. & Foster, S. (2005). Emerging Issues in Location Based Tourism Systems. Proceedings of the International Conference on Mobile Business (ICMB'05) (pp. 75- 81). IEEE Computer Society. Hepp, M. (2006). Semantic Web and semantic Web services: father and son or indivisible twins? Internet Computing, IEEE. 10(2). 85- 88. Hepp, M., Siorpaes, K. & Bachlechner, D. (2006). Towards the Semantic Web in E-Tourism: Can Annotation Do the Trick? In Proc. of 14th European Conf. on Information System (ECIS 2006), June 12–14, 2006, Gothenburg, Sweden. Heung, V.C.S. (2003). Internet usage by international travellers: reasons and barriers. International Journal of Contemporary Hospitality Management, 15 (7), 370-378.

17

Horridge, M., Knublauch, H., Rector, A., Stevens, R. & Wroe, C. (2004). A Practical Guide To Building OWL Ontologies Using The Protege-OWL Plugin and CO-ODE Tools Edition 1.0. The University of Manchester, August 2004. [Online]. Available: http://protege.stanford.edu/publications/ontology_development/ontology101.html Jakkilinki, R., Georgievski, M. & Sharda, N. (2007). Connecting Destinations with an OntologyBased e-Tourism Planner. In Sigala, M., Mich, L. & Murphy, J. (Eds.): Information and Communication Technologies in Tourism 2007, Proceedings of the International Conference in Ljubljana, Slovenia, 2007. (pp. 21-31). Springer Vienna. Kanellopoulos, D. & Kotsiantis, S. (2006). Towards Intelligent Wireless Web Services for Tourism. IJCSNS International Journal of Computer Science and Network Security. 6 (7). 83-90. Kanellopoulos, D.,Kotsiantis, S. & Pintelas, P. (2006), Intelligent Knowledge Management for the Travel Domain , GESTS International Transactions on Computer Science and Engineering. 30(1). 95-106. Popov, B., Kiryakov,A., Ognyanoff, D.,Manov, D. & Kirilov, A. (2004). KIM - A Semantic Platform For Information Extraction and Retrieval. Journal of Natural Language Engineering, Cambridge University Press. 10 (3-4). 375-392. Roman D., Keller, U., Lausen, H., Bruijn J. D., Lara, R., Stollberg, M., Polleres, A., Feier, C.,Bussler, C. & Fensel, D. (2005). Web Service Modeling Ontology. Applied Ontology. 1(1). 77 - 106. Shadbolt, N., Berners-Lee T. & Hall, W. (2006). The Semantic Web Revisited. IEEE Intelligent Systems. 21(3). 96-101.

18

Smith, C. F. & Alesso, H. P. (2005). Developing Semantic Web Services. A K Peters, Ltd. Stojanovic, LJ., Stojanovic N. & Volz, R. (2002). Migrating data-intensive Web sites into the Semantic Web. Proceedings of the 2002 ACM symposium on Applied computing, Madrid, Spain, ACM Press. 1100-1107. TERMS AND DEFINITIONS Web portal: a Web site or service that offers a broad array of resources and services, such as email, forums, search engines, and on-line shopping malls. Ontology: a controlled vocabulary that describes objects and the relations between them in a formal way, and has a grammar for using the vocabulary terms to express something meaningful within a specified domain of interest. Intelligent agents: software elements which help the user find information of specific interest to him/her without their explicit assistance. Intelligent reasoning: the act of using reason to derive a conclusion from certain premises using a given methodology. Semantic Web services: self-contained, self-describing, semantically marked-up software resources that can be published, discovered, composed and executed across the Web in a taskdriven semi-automatic way. Dynamic packaging: the combination of different travel components, bundled and priced in real time, in response to the requests of the consumer or booking agent. Location Based Services: services that provide context-sensitive information based on the mobile user’s location.

19

Web Service Modeling Ontology (WSMO): a data model that provides the conceptual underpinning and a formal language for semantically describing all relevant aspects of Web services in order to facilitate the automation of discovering, combining and invoking electronic services over the Web. OWL-based Web service ontology (OWL-S): an ontology which supplies Web service providers with a core set of constructs for describing the properties and capabilities of their Web services in unambiguous, computer-interpretable form.

20