شخصی سازی ارائه محتوا و انطباق رسانه ها در وب سایت های گردشگری با استفاده از روش دلفی فازی و نقشه های فازی شناختی

Web sites become more powerful when they can adjust to their users’ needs. Web personalisation refers to adapting both the content and the presentation of web sites, so that to deliver the maximum effect to the user in the most appropriate way. A main objective of web personalisation is to adapt the presentation of the web content in a manner that increases the user’s perceived quality. This paper focuses on the applicability of fuzzy logic techniques to content presentation and media adaptation. More specifically, it applies Fuzzy Delphi Method (FDM) and Fuzzy Cognitive Maps (FCMs) in order to highlight the services features that are most preferred by the customer and to adapt presentation media and layout. Fuzzy logic is utilised to deal with the subjectivity inherent in web design choices and in customers’ perception of services priorities. FDM is used to capture the experts’ knowledge regarding media adaptation with respect to hotel service quality. A prototype that adapts the web site presentation according to customer preferences has been developed and evaluated.

Web adaptation and personalisation is a major development towards web services that adapt to users’ requirements and increase their satisfaction and loyalty. Adaptation is therefore suggested as a major determinant of e-service quality (Kaynama and Black, 2000 and Murugesan and Ramanathan, 2001). The adaptation of web systems includes: adaptive content and recommendation, adaptive navigation and search and adaptive presentation (De Virgilio, Torlone, & Houben, 2006). Web personalisation applies to ‘what to offer’, i.e. what content to provide, and ‘how to offer it’, i.e. in what style and format to present the content. Web adaptation can be thought of as a two-task process: content adaptation, which aims at selecting the content that is most relevant to user needs and priorities, and content presentation adaptation, which aims at deciding how to most effectively present the selected content to the user (Bunt, Carenini, & Conati, 2007). Main objectives of web personalisation include (Murugesan & Ramanathan, 2001): • To personalise web interface and navigation • To provide personalised information content • To present web site content in personalised format and layout • To suggest special offers to selected online customers • To collect and analyse data and to investigate e-customer behaviour • To develop a ‘human face’ for the web. The term adaptive web site describes sites that manually, semi-automatically or automatically adapt their organisation and presentation according to the visitor’s behaviour and characteristics (profile) (Perkowitz & Etzioni, 2000). Many types of online businesses, such as portals and e-catalogues employ adaptive web site technologies. They provide for example, customised local weather reports, favourite links and special offers. Some of them introduce new pages (content), insert, delete or even highlight hyperlinks (navigation) and change the format of specific points in a page (presentation). Some create a different version of the site for each user or for different user groups (Perkowitz & Etzioni, 2000). Another kind of adaptation is reforming and organising the content and its structure. The degree of adaptation depends on the level of automation. Some systems try to predict the user’s next page or even his/her interests (Perkowitz & Etzioni, 2000). Personalisation can be offline, for example, when personalised newsletters and e-mails are sent to customers, or online, in which case data analysis and recommendations are carried out in real time. Personalisation in terms of its sophistication can be shallow (for example, addressing a customer by his/her name) or it can be deep, where personalised content, navigation and presentation can provide the foundation for one-to-one marketing. Web personalisation involves three main steps (Murugesan & Ramanathan, 2001): (1) User behaviour and profile data collection. (2) Analysis of user data. (3) Recommendation of personalised content and personalised presentation. This paper focuses on content presentation personalisation and media adaptation. It considers four media types (text, icon, photo, and video) and determines the degree of usage for each media type in presenting hotel e-services related content. The paper also aims at adapting the presentation of e-tourism web sites by taking into consideration the preferences of hotel service as perceived by individual customers. It investigates the applicability of fuzzy logic in presentation personalisation and applies the Fuzzy Delphi Method and Fuzzy Cognitive Maps (FCMs) for selecting the most appropriate media, in order to present selected content and to develop adaptive web sites. A prototype system, the Fuzzy Web Site Adaptation System (FWSAS) is also presented. FWSAS draws on experts’ knowledge, regarding the appropriateness of media types, in presenting individual customer hotel services preferences, in order to customise the web sites presentation, according to customers’ priorities. FWSAS has been tested for its effectiveness, by analysing the responses of a group of 36 users. The structure of this paper is as follows: The next section discusses approaches to media selection for presentation personalisation. Next, the paper discusses the adopted methodology and the modelling approach followed by the section, which describes the fuzzy algorithm for personalisation and the architecture of the prototype system. The paper continues with an illustrative example and the evaluation of the prototype. Finally, the last section reviews the benefits of the approach and suggests directions for further research.

The proposed approach for intelligent adaptive web sites benefits from the use of fuzzy logic, in the sense that user perceptions of service quality are automatically translated to features of an adaptive web site. Customers find it easier to describe their service quality preferences in qualitative terms but even so, they do not apply such preferences rationally (due to biases and cognitive limitations) when assessing the overall quality of an offered service. Customers therefore need to have the most valued quality features highlighted in the service representation and the less relevant ones suppressed or even omitted. Although the system described here focuses on media adaptation and web page layout, the same principles can be applied to other types of content presentation management, i.e. where fuzzy content filtering techniques can be applied based on a customer profile which is also fuzzily described. The prototypes’ evaluation results are promising and further research is suggested so that fuzzy logic and fuzzy measures can be incorporated into more design factors (van Iwaarden, van der Wiele, Ball, & Millen, 2004) such as location, size, used colour, response time, and other usability factors. When manual completion of a customer profile is not possible, automatic profile generation from Customer Relationship Management systems (CRM) or other customer logged behaviour data and patterns can be used as alternatives. In conclusion, the use of fuzzy logic opens up opportunities to improve presentation personalisation.