Behavioural - based modelling and analysis of Navigation Patterns across Information Networks

                                                                                 UDC: 004.83:519.21 

Behavioural – based modelling and analysis of Navigation Patterns across Information Networks Vesna Kumbaroska1 and Pece Mitrevski2 1 University of information science and technology, St. Paul the Apostle, Partizanska bb, 6000 Ohrid, Macedonia vesna.gega@uist.edu.mk 2 University St. Clement of Ohrid, Faculty of information and communication technologies, Partizanska bb, 7000 Bitola, Macedonia pece.mitrevski@fikt.uklo.edu.mk Abstract. Navigation behaviour can be considered as one of the most crucial aspects of user behaviour in an electronic commerce environment, which is very good indicator of user’s interests either in the process of browsing or purchasing. Revealing user navigation patterns is very helpful in finding out a way for increasing sale, turning the most browsers into buyers, keeping costumer’s attention, loyalty, adjusting and improving the interface in order to boost the user experience and interaction with the system. In this regard, this research has identified the most common user navigation patterns across information networks, illustrated through the example of an electronic bookstore. A behavioural-based model that provides profound knowledge about the processes of navigation is proposed, specifically examined for different types of users, automatically identified and clustered into two clusters according to their navigational behaviour. The developed model is based on stochastic modelling using the concept of Generalized Stochastic Petri Nets which complex solution relies on Continuous Time Markov Chain. As a result, calculation of several performance measures is performed, such as: expected time spent in a transient tangible marking, cumulative sojourn time spent in a transient tangible marking, total number of visits in a transient tangible marking etc. Keywords: user behaviour, navigation behavior patterns, Generalized Stochastic Petri Net, Continuous Time Markov Chain. 1. Introduction The growing popularity of electronic commerce leads to an increased number of buyers and very frequent online purchasing. Every day, new online shopping standards are established, based on changed and grown user demands and expectations which implies a need to retrieve more precise results in the process of searching, accurate recommendations, better design oriented to the user, and adequate personalization. This requires a profound understanding of the interaction, i.e., the interface, on one hand, and Behavioural – based modelling and analysis of Navigation Patterns across Information Network 61

the user behaviour, on the other hand, which is extremely important to create new strategies. Navigation behaviour can be considered as one of the most important aspects of user behaviour in an electronic commerce environment, which also is very good indicator of user’s interests. In this direction, the main idea of our work is to identify the most frequent navigation patterns and show how Petri Nets (PN), as transition based models, can be applied in modelling user navigation behaviour. In this work, we want to outline some intrinsic details of Generalized Stochastic Petri Nets (GSPN) application for describing user navigation patterns, which actually take some time to execute (perform).
The rest of the paper is organized as follows. In Section 2 we present related work in this field. Section 3 represents an overview of the Petri Net formalism and our contribution. A case study is presented in Section 4. In the last section, we give some conclusions and steps for future work. 2. Related Work To build an effective user navigation behaviour model means to develop an accurate predictive mathematical model of the user behaviour. Usually, the models are based on log data, collected in a period of time [16, 20], which structure is complex, but it carries an important source of information about user behaviour. In order to study log data and to gain knowledge of how users navigate, statistical analysis and application of data mining techniques need to be performed. Usually, the emphasis is placed on developing models for: discovering user search or navigation patterns [24], predicting and proposing future user actions [7] and personalization based on user behaviour [2, 12, 21]. In order to discover common user navigation behaviour patterns and also common sequences of transitions, a data mining approach is used by [19], based on a particular case of log data, taking into account the duration of the website visits. Also, a data mining approach, but in combination with semi-Markov process in discrete time, is applied in [10], in order to understand and describe the user behaviour. They propose an algorithm for obtaining a transition probability matr

…(Full text truncated)…