Automatic Identification of Unique Conference Names using Rule-based System

An academic conference is one popular meeting platform where researchers could present and discuss their works in a particular research topic of interests. One of the well-known databases including conference publications is Scopus, which is also used in this paper as a case study. Most of the time, statistical data related to conference publications including citations is used in many real-world analyzes such as constructing conference ranking systems, and conference quality assessment. The statistical records and analysis on individual conferences are then necessary. However, a main difficulty of the automatic analysis is names of each conference are usually recorded and referred in the database differently. They may include different styles of year tags, and decorate words such as IEEE, ACM, and proceeding. Therefore, this paper proposes a novel rule-based approach to eliminate duplicated conference names and re-assigned them with unique pseudo IDs which can be used for any further analysis. The proposed method is validated on ERA2010 and CORE2018 with relative errors of 0.7% and 0.4% respectively. Then, it is applied on Scopus 2006-2010, Scopus 2010-2014, and Scopus 2014-2018, with reported duplicated conference names of 38.9%, 42.6%, and 36.2% respectively.