A method for identifying clusters in sets of interlinking Web spaces

摘要

A technique is presented for the identification of patterns from the links between large Web spaces and is applied to data concerning the interlinking of university Web sites in fifteen European countries. This is based upon a procedure for normalising the data so that it can be analysed using standard multivariale statistical techniques and is less susceptible to individual outliers than standard methods. The approach was successfully able to identify clusters of European countries based upon data for their universities' interlinking patterns. For example, the northern countries were differentiated from the southern with this method.