Towards a Time-based Approach for Author Co-citation Analysis

The Author co-Citation Analysis (ACA) is a widely used statistical technique for mining information about those authors who publish in related research domains. The existing ACA technique generates author clusters by initially defining the co-citation count. Co-citation count between authors of two different research papers is defined as the number of times these authors are cited together by a set of source papers. The technique used to determine the co-citation count needs to be effective as it greatly influences the obtained author clusters. This paper presents an enhanced ACA that utilizes a novel co-citation counting technique. The enhanced ACA technique takes into consideration the research papers referred to in the source paper, the papers that have cited the source paper, and their publication year. Experimental results obtained indicate that the author clusters produced, comprise primarily of active researchers having published in the recent time period, specified in years. In this study, we have assumed that active researchers are those who have published in or after the year 2000. The proposed Time based ACA (TACA) technique uses a real time data set consisting of papers collected from ACM’s Transaction on Database Systems (TODS) journal from the year 2006-2009. The average precision of the proposed technique is found to be around 93%, when evaluated against the benchmark ACM Computing Classification System (CCS).