Analyze user behavior patterns on academic search engines

Nowadays, academic search engines have grown rapidly. Thus, understanding the users’ information‐seeking patterns has become one of the most important research topics. That is why by examining user interaction logs, developers can discover user behavior patterns to determine who they are and what they tend to do. Consequently, they can get guidance to design better academic search engines and improve their performance. In this paper, we analyze search engine users’ logs gathered from the search engine of the Iran scientific information database (Ganj). The users are clustered into three distinct groups: fast surfing, broad scanning, and deep-diving, using the K-means clustering algorithm. After that, we investigate the frequent sequences of behavior patterns and the networks of search keywords for each cluster separately. The results show that users with similar information‐seeking patterns have similar sequences of behavior patterns. The findings can help the developers of academic search engines and policymakers to identify users' needs and priorities and make better decisions

Keywords:

academic search engines , log mining , clustering algorithm , behavior analysis , information‐seeking patterns