Інформаційна система ієрархічної категоризації науковців за відкритими даними з Google Scholar. Частина 1. Інтелектуальний аналіз даних

Abstract

In the work, a method for multi-label hierarchical classification of researchers based on their interests in Google Scholar is presented. The method allows to automatically classify a researcher to a few research specialities and domains that fit for him the most in ANZSRC system. The first research result is the model to assess similarity of research specialities based on statistic of publications in Dimensions and Jaccard’s index. The second research result is improved categorization method that: a) in addition to interests the pair of interests of a researcher is used; b) research specialities reduction uses specialities’ similarity. Based on research results the information system was created.