Intelligent Interaction: A Case Study of Web Page Prediction Andrea Ba i and Andrina Grani * Siemens d.d., Siemens IT Solutions and Services, Put brodarice 6, 21000 Split, Croatia E-mail: andrea.bacic@siemens.com *Faculty of Science, University of Split, Nikole Tesle 12, 21000 Split, Croatia E-mail: andrina.granic@pmfst.hr Abstract. In this paper preliminary work in the area of web page prediction is presented. The designed and implemented prototype offers personalized interaction by predicting the user's behaviour from previous web browsing history. Those predictions are afterwards used to simplify the user's future interactions. Rather simple and feasible prototype enhancements are offered and discussed. Its simplicity and effectiveness makes it potentially useful for widespread application. Keywords. web browsing, web page prediction, personalized interaction 1. Introduction Human-Computer Interaction (HCI) research places an individual as the focus of all theoretical and practical advances, stressing the importance to design technologies for human needs. Intelligent user interfaces (IUIs) have been recommended as a means of making systems individualized or personalized, thus enhancing the systems flexibility and attractiveness [2; 11]. Specific applications of intelligence to areas as diverse as intelligent hypermedia, recommender systems, intelligent filtering, explanation systems, intelligent help and intelligent tutoring could be identified. Thus, the intelligence in the interface of such a system can make the system adapt to the needs of different users, can take initiative and make suggestions to the user, can learn new concepts and techniques, can provide explanation of its actions or can predict the next action of the user cf. [3; 14]. The ability to predict the user's next action allows the system to envisage the user's needs and to adapt to and improve upon the user's work, aiding the human-computer interaction process. In this paper we present preliminary work in the prediction of the user's next requested web page. The designed and implemented prototype offers personalized interaction by predicting the user's behaviour from her/his previous web browsing history and then uses these predictions to simplify her/his future interactions with the browser. The paper is organized as follows. In Section 2 we present a short outline of research in the area of user actions prediction. Section 3 is dedicated to describing used technology, the architecture of the system and the developed prototype for web page prediction. In Section 4 we provide discussion, suggestions for prototype improvement along with directions for further work. 2. Related research and motivation Various systems able to predict the user's next action are briefly presented in the following. Early work in command line prediction includes Greenberg's (1988) work on dataset’s, still usable even today [9], Darragh, Witten and James's (1990) Reactive Keyboard [5] and Greenberg, Darragh, Maulsby and Witten's (1991) Predictive Interfaces [10]. More recently, Davison and Hirsh (1997) have modified the UNIX shell in order to memorize a command history and subsequently (by means of different methods) calculated the probability of a certain command input [6; 7]. Based on their work is the research of Korvemaker and Greiner (2000) who have studied the prediction of the next UNIX command applying diverse statistical methods [13]. In addition to the command input, they have also attempted to predict the command parameters. Jacobs and Blockeel's (2003) have tried to improve Davison and Hirsh's command prediction by employing a time component [12]. That is, they assumed that the time of the command input should be relevant for next UNIX command prediction. Several authors have based their research on predicting the next requested web site. Zukerman, Albrecht and Nicholson (1999) have addressed the ways for minimizing waiting for a requested web page by predicting which web page the user will ask for next [17]. Based on their work is the research of Friaz-Martinez and Karamcheti (2002), who have enhanced their predicting model 287 Proceedings of the ITI 2009 31 st Int. Conf. on Information Technology Interfaces, June 22-25, 2009, Cavtat, Croatia