A Method to Support Web Evolution by Modeling Static Structure and Dynamic Behavior

Software evolution is an important activity in software life cycle. The Web application is very different from traditional software in design and composition. It needs maintained more frequently. So the automatic evolution support is a critical part in Web engineering. This paper uses Markov chain to model Web userspsila browser behavior by statistic analysis of Web log. And transition probability is used as weight in model. By it a model of browse centrality is built. Contrasting it with static model of structure centrality we propose an algorithm based on rules about automatic evolution of Web link structure.