Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10355164 | Information Processing & Management | 2005 | 16 Pages |
Abstract
Previous studies of academic web interlinking have tended to hypothesise that the relationship between the research of a university and links to or from its web site should follow a linear trend, yet the typical distribution of web data, in general, seems to be a non-linear power law. This paper assesses whether a linear trend or a power law is the most appropriate method with which to model the relationship between research and web site size or outlinks. Following linear regression, analysis of the confidence intervals for the logarithmic graphs, and analysis of the outliers, the results suggest that a linear trend is more appropriate than a non-linear power law.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science Applications
Authors
Nigel Payne, Mike Thelwall,