Degree of Similarity of Root Trees

Adaptive User Interfaces (UI) provide better user experience as users a receive personalized presentation. These UIs heavily rely on contextual data. Context helps the application to recognize user needs and thus adjust the UI. First time user receives a generalized experience; however, as the user uses the application more often it gathers lots of contextual data, such as the history of actions. This allows to statistically classify user in a user cluster and based on that adapt the UI presentation. This paper considers methods to find a measure of similarity of graphs to support adaptive UIs. To achieve this, it considers rooted trees. It states known approaches, which could be used for calculation of this measure. It focuses on the Simhash algorithm and describes its implementation in the SimCom experimental comparative application. Its results show that Simhash can be used for comparing the rooted trees. The main aim of this paper is to show novel view on how to use graph algorithms and clustering of trees into adaptive application structure.