hu-index: a unified index to quantify individuals across disciplines

Nowadays scientific evaluation is becoming increasingly important and necessary in many cases, such as faculty hiring, funding and promotion. Among existing evaluation metrics for individual performance, h-index is the most famous indicator and achieves a prominent role since its publication. However, h-index is inapplicable to comparing individuals from different scientific disciplines, primarily because it cannot handle the huge difference in collaboration habits and citation practices across disciplines. Such a shortcoming of h-index is rooted in its arbitrary definition, comparing two quantities with quite different scales, i.e., citation count for scientific impact and publication count for productivity. To combat this problem, we propose a new evaluation measure, h(u)-index, which unifies citation count and publication count into the same scale. We theoretically analyze the relationship between h(u)-index, h-index and other variants of h-index. We also study the behavior of h(u)-index in empirical cases and researcher ranking tasks. Experimental results demonstrate that h(u)-index has superior performance than h-index and achieves a better comparison of individuals across disciplines.

Automated summary

Scientific evaluation is increasingly important, with the h-index being a famous indicator for individual performance. However, it lacks the ability to compare individuals across different scientific disciplines. To address this issue, a new evaluation measure, h(u)-index, has been proposed to provide a better comparison across disciplines.