Topic modeling of researchers based on their interests from Google Scholar

The article proposes an algorithm for topic modeling of researchers based on their interests from Google Scholar profiles. The algorithm uses the set of fields of research from research classification system ANZSRC. An information resource for topic modeling is a corpus of categorized publications from Dimensions. Interests from researchers’ profiles are used as search queries to Dimensions that outputs distributions of documents over categories. To reduce information noise these distributions are taken through a few stages of processing. The article also compares the results of topic modeling based on interests from Google Scholar profiles and based on a categorized list of publications from Dimensions. The comparison is done using modified Czekanowski metric that takes into account the similarity between categories. The results of comparing the topic modeling outputs based on different information sources show a good match.