Research specialties’ kinship level identification based on data from Dimensions

Abstract

Knowledge about research specialties’ kinship level is needed for solving such problems as: improving current research classification system; detecting similar scientific and educational institutions to set up cooperative relations or perform their reorganization; automatic reviewer assignment for peer reviewing PhD-thesis, papers, grant proposals etc. In this paper research specialties’ kinship level is identified according to Australian and New Zealand standard research classification ANZC-RC-2020. The identification is done using information system Dimensions by analyzing 33.8 million publications for 2019-2023. The level of kinship is assessed by Jaccard index as the ratio of two specialties common publications’ number to the total number of publications in these specialties. It is found, that from 14535 possible pairs of specialties only 131 pairs have significant kinship with Jaccard index greater than 0.05. For 20 pairs among them the kinship level is high, and for 61 pairs – average.