Entwicklung und Test einer logfilebasierten Metrik zur Analyse von Website Entries am Beispiel einer akademischen Universitäts-Website

Abstract

Web log files record user transactions on web servers and offer due to their extent, their properties and potential an excellent investigation field for contemporary information and online behaviour studies. The empirical, explorative investigation from the fields Web Mining, Webometrics and Logfile Analysis introduces new aspects and analysis possibilities for web log data. The study develops and tests a quantitative, non-reactive measure method (log metric "Web Entry Factors"), that allows statements about the accessibility and visibility of highly frequented entry points of a web site. The three distinguishable web navigation types "navigation about search engines", "navigation about backlinks" and "direct navigation" are focused. The study integrates a classification scheme for web pages as well as the prominent external parameter PageRank from the today most important search engine Google. Investigation subject are Web log files of two complete age-groups (in 2000 and 2002) of the web server of the Institute for Library Science at the Humboldt University to Berlin (http://www.ib.hu-berlin.de/), as well as the 100 most frequently used entry pages of this academic university web site.This publication goes back to the master thesis in Library and Information Science (Master of Arts, M.A.) at Humboldt-Universit�t in Berlin.