Proposal of Research Activity Classication Based on Time-series Bibliometrics

Abstract

Bibliometrics, such as the number of papers cited and frequency, are often used to compare researchers based on specific criteria. The criteria, however, are different in each research domain and are set by empirical laws. Moreover, there are arguments, such as that the simple sum of the metric values works to the advantage of the elders. Therefore, this paper attempts to constitute features from the time series data of bibliometrics and then classify the researchers accordingly. In detail, time series patterns are extracted from a large amount of bibliographic datasets, and then a model to classify whether the researchers are "distinguished" or not is created by using machine learning techniques. The experiments achieved F-measures of more than 80% in the classification of 114 researchers in two research domains based on the datasets of the Japan Science and Technology Agency and Elsevier's Scopus. In the future, we will conduct verification on a number of researchers in several domains, and then we will make use of discovering "distinguished" researchers who are not yet widely known.