User Evaluation of Web Clustering Search Engines

This study provides user evaluation of clustering and ranked list search engines by comparing their search efficiency, search effectiveness, and users' satisfaction level. The methods include experiment, observation, questionnaire, interview, and search log analysis. The results show that the ranked list search engine performs better in users' search speed, relevant pages retrieved, and the satisfaction level. On the other hand, the clustering search engine has the merits of demonstrating important concepts, providing clues for diverse thinking, and helping to reduce the information overload. Meanwhile, users' acceptance of individual clusters is higher than that of the cluster scheme. The paper also provides some discussions and suggestions on improving the clustering search engines.