出版社：Consejo Superior de Investigaciones Científicas
摘要：This study analyzes how users perform information retrieval tasks when introducing queries to the Hispanic Digital Library. Clusters of users are differentiated based on their distinct information behavior. The study used the log files collected by the server over a year and different possible clustering algorithms are compared. The k-means algorithm is found to be a suitable clustering method for the analysis of large log files from digital libraries. In the case of the Hispanic Digital Library the results show three clusters of users and the characteristic information behavior of each group is described.
关键词：Clustering;k-means algorithm;digital libraries;log files;Transaction Log Analysis (TLA);Biblioteca Digital Hispánica;Agrupamiento;algoritmo k-means;bibliotecas digitales;ficheros log;análisis de ficheros de transacciones;Biblioteca Digital Hispánica
