Decision analysis

Expert assignment method based on similar document retrieval from large text collections

D. V. Zubarev^a, I. V. Sochenkov^ab, I. A. Tikhomirov^a, O. G. Grigoriev<sup>a

a</sup> Federal Research Center "Computer Science and Control" of Russian Academy of Sciences, Moscow, Russia
^b Skolkovo Institute of Science and Technology, Moscow, Russia

Abstract: The article is devoted to the task of expert assignment. The article provides an overview of methods, which are currently used to solve this task. We discuss the main problems of those methods and propose to leverage large collections of documents that are authored by the experts. The article describes a basic method for searching and ranking of experts for a given document, using similar document retrieval. For the evaluation of the proposed method we use private corpus of applications for a grant. Experimental studies show that the more documents are available that are authored by experts, the better recall becomes. In conclusion we discuss current limitations of proposed method and describe future work to use more features from texts, such as bibliography, co-authors information etc.

Keywords: scientific expertise, expert assignment, analysis of unstructured data, text analysis, similar document retrieval.

DOI: 10.14357/20718594190206

Bibliographic databases:

•