Current online businesses usually contain supporting tools for an instant acquiring offered services and goods. Electronic web site solutions that guide a potential user to a successful finalization of browsing activities, i.e. to purchase finalization play an important role in trading. The proposed in the paper approach presents a model aiming at an assistance in transaction finalization using browsing activity data and personal information. The model is capable to discover missing personal data and to use forecasted values in the research. Results of the applied machine learning techniques are compared while using the whole set of the collected browsing activity data and the data with a reduced number of dimensions.