(Wydawnictwo Uniwersytetu Rzeszowskiego, 2015) Budnikas, Germanas
The goal of the research was to create a user’s online behaviour analysis model. The goal was achieved by accomplishing a set of phases. A general user’s behaviour scenario at a web page of category „Consumer Goods and Services” with respect to Open Directory Project was formalised using Business Process Modeling Notation.
Needed data to represent behaviour was defined and include user’s IP address, task id and time moment to be passed using Google Analytics TrackEvent() function while user browses a web page. Historical data about user behaviour are used for an artificial neuron training. After the training, most important factors that affect transaction finalisation (e.g. service order) can be defined. While monitoring actual on-line user’s behaviour the developed artificial neuron model permits the estimation of a behaviour outcome – would it be finalised or not. In a later case, it is foreseen an interaction of a web site, which can suggest visiting pages that has the most impact level for transaction finalisation. In such a way, competitiveness of an enterprise is achieved.
