Weighting Factor Mechanism of Uncited Papers to Improve the Fairness of RA-index based on Particle Swarm Optimization

摘要

Today, Clarivate Analysis (WoS), Scopus and Google Scholar use the H-Index method to figure the profile of the researchers. The advantages of the H-index are simple mathematic application and can be implemented in many areas. However, the H-index also has weaknesses, such as less sensitive to measure the impact of productive and perfectionist groups of researchers. Many proposals tried to improve the H-index method, such as the well-known G-index. H-index and G-index did not count the number of citations, which is the citations below the H-index value and do not count uncited papers to contribute the index value. This study proposes weighting on uncited papers and the number of citations below the H-index value (lower-h-tail area), as the data to be considered to be measured as an impact to differentiate researchers. The weighting is related to a new method of calculating the impact of the researcher from previous work, namely RA-index. The RA-Index method accommodates productive researchers based on Lotka's law and the Fairness Jain theory. Particle swarm optimization (PSO) method is used to optimize the weighting of the uncited papers. Weighted values for uncited papers and the number citation of papers with citation value of the H-index values obtained is 0.52. The result of discrimination test from the weighted value manually of 0.50 is 0.09, while the weighting of optimization results obtained is 0.08. From these results, it is concluded that the PSO method is feasible to be used in optimizing the weighting of the RA-Index method. The comparison results showed that the RA-Index is fairer than its competitors. The result of fairness calculation for RA-index (optimized) has fairness value of 0.92 higher than RA-index (not optimized) that has fairness value of 0.91.