1. Introduction In this paper, we discuss a graphical presentation of the similarity relationship between web pages. There have been several studies on the notion of similarity between web pages, but the similarity of web pages has not been widely utilized by users. We attribute this to the lack of an appropriate interface to navigate similar pages. We propose using a hyperbolic tree to represent the similarity relationship to allow users to see how the results of a search query are related to each other.
Currently Google provides the user with a simple search interface. When search results are presented to the user, each result also has a “similar pages” link that will display similar results. We build our interface on top of Google to take advantage of this similarity search feature. We present the results in a hyperbolic tree, using an interface that allows the user to navigate through the tree of similar pages.
2. GoogleAPI The googleapi provides a convenient way of accessing the Google search engine. It allows us to send queries, and retrieve the results through a well defined interface, saving us the task of parsing search results from a web page. The interface does not directly support searching for similar pages, it allows us to send queries formatted as “related:www.somewebsite.com” to fetch similar pages.