Co-occurrence Analysis of Access Log of Institutional Repository

Institutional repository is playing an important role to guarantee open access to research outputs by self archiving. However, the number of the items in most institutional repositories is extremely fewer than that of the total research outputs produced in the institute. One of the reasons is that most researchers have no incentive to register their research outputs, simply because the effectiveness of registration to institutional repository is not clear. The authors are constructing a feedback system for researchers who register their research outputs to institutional repository. In this paper, they focus on access log analysis to discover meaningful knowledge on when, how, and why the items are accessed. The knowledge from the access log can utilized also for recommendation of items for users (readers) of the institutional repository. This paper shows some results of co-occurrence analysis for access log of the institutional repository of Kyushu University, and shows some ideas of advanced analysis to obtain meaningful knowledge. keyword: Institutional repository, access log, recommendation, co-occurrence, visualization, formal concept analysis.