Possible Ways of Applying Citations Network Analysis to a Scientific Writing Assistant
Development of linguistic technologies gave rise to a new type of tools for academic writing, which use natural language processing and heuristics to help authors write scientific papers. In our contribution we present a new function “advise a paper to read” and the way it could be implemented. We discuss a possibility of using different centrality metrics and test their application in 50 cases created from 50 top cited articles of the Engineering domain. For each case we created a citation network graph based on the results of a search query in the Web of Knowledge by Thomson Reuters, using adjusted authors key phrases, and compared the results of applying the centrality metrics with the actual reference list presented in each article.