Design And Implementation Of User Behavior Analysis System Based On Interactive Process

With the amount of website users increasing, the administrators are eager to know the users’ behavior in order to help manage the site and provide decision support.Therefore the most urgent thing is to identify the "law" of the interaction between users and webpages. Operations at the site can be traced, sessions can be event traces. It is the key to mining actual interaction process from a mass of operation records. The mining results will reflect if the majority users adapted to the sites and also indirectly reflect the user’s operating habits.Currently the study on interaction process in the field of webpage users behavior analysis still stops at basic solutions such as a page-based thermodynamic diagram, pages access sequence diagram, click-stream model. Without model analysis results, the lack of comparison with reference model and unable to complete performance analysis have become the main disadvantages except that the solutions are unable to refine to event.The core requirement is mining interaction process in this thesis. Firstly, author proposes an extension to the flowchart to express reference model better. Secondly, using the a algorithm to excavate the actual interaction model from the event tracks. Then analyzes the differences and calculate the similarity by comparing with the reference model. Finally, counts the performance quantitatively of real interactive process. Besides above, the system can also realize three basic statistics such as site traffic analysis, content efficiency analysis and user analysis to meet more needs.The topic is an actual project that author participated in as an intern. Author is responsible for the functional design and development in this project. After the analysis of requirements, introduce the algorithm design, the technical architecture and the functions in detail.Finally, Author demonstrate the feasibility of mining algorithm by a real online website of tickets purchasing and prove certain value of research and application.