Identifying a better measure of relatedness for mapping science

Measuring the relatedness between bibliometric units (journals, documents, authors or words) is a central task in bibliometric analysis. Relatedness measures are used for many different tasks, among them the generating of maps, or visual pictures, showing the relationship between all items from these data. Despite the importance of these tasks, there has been little written on how to quantitatively evaluate the accuracy of relatedness measures or the resulting maps. We propose a new framework for assessing the performance of relatedness measures and visualization algorithms that contains four factors: accuracy, coverage, scalability, and robustness. This method was applied to ten measures of journal-journal relatedness to determine the best measure. The ten relatedness measures were then used as inputs to a visualization algorithm to create an additional ten measures of journal-journal relatedness based on the distances between pairs of journals in two-dimensional space. This second step allows us to determine robustness (i.e., which measure remains best after dimension reduction). Results show that, for low coverage (under 50%), the Pearson correlation is the most accurate raw relatedness measure. However, the best overall measure, both at high coverage, and after dimension reduction, is the cosine index or a modified cosine index. Results also showed that the visualization algorithm increased local accuracy for most measures. Possible reasons for this counterintuitive finding are discussed.