A semantic self-organising webpage-ranking algorithm using computational geometry across different knowledge domains

In this paper we introduce a method for Web page-ranking, based on computational geometry to evaluate and test by examples, order relationships among web pages belonging to different knowledge domains. The goal is, through an organising procedure, to learn from these examples a real-valued ranking function that induces ranking via a convexity feature. We consider the problem of self-organising learning from numerical data to be represented by a well-fitted convex polygon procedure, in which the vertices correspond to descriptors representing domains of web pages. Results and Statistical evaluation of procedure show that the proposed method may be characterised as accurate.