Association Rules Mining for Academic Cooperation Based on Time Extension and Duration Accumulation

Academic relations originated from academic activities are of particular temporal features. The academic activities which happen more recently, last longer period and have more occurrence will be more possible to form valuable academic relationships and more meaningful to be analyzed. For the characteristics of the accumulated time and vague boundaries of valid time with academic cooperation, the strategy of time extension and duration accumulation is introduced into the algorithm of progressive partition association rules mining. In the proposed method, the time attribute of academic activities transaction records are mapped and extended on time series, and the duration period is marked according to time partition. Through progressive partition processing, the academic association rules mining with continuous time accumulation is realized. This paper presents basic principle and realization process in details, and takes information about research projects and co-published papers as cases to mine academic relationships. The experimental comparison and results analysis show the effectiveness of the proposed algorithm.