Predicting on-line task completion with clickstream complexity measures: A graph-based approach

A model that uses navigational complexity to classify on-line shopping sessions and information-search tasks as successful or unsuccessful is described. A graph-based approach to representing clickstream is employed to capture the complexity. A total of 485 individual goal-oriented shopping sessions on five different Web sites were analyzed. Ten complexity metrics were studied in order to select the ones that can predict the outcome (i.e., success or failure) of a goal-oriented navigational session. The practical applications to e-business, including the implementation of the proposed model, are discussed.