期刊
JOURNAL OF INFORMETRICS
卷 15, 期 2, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.joi.2021.101130
关键词
Junior researcher; Research performance; Deep learning; Linear regression
资金
- Maranhao Research Foundation (FAPEMA) [BD-08792/17]
This article introduces a model that predicts the stable Q values of junior researchers using data from their initial years of publication, and compares the accuracy of a deep model and linear regression model. The results show that both models' predicted values are more reliable than Q values calculated by the Q model itself using only data from the first five years of publication. This approach is also robust in dealing with citation bias inflation.
A researcher's Q denotes their ability in scientific research as a real number. Due to their short presence in the academic environment, junior researchers have unstable Q values. This article aims to present a model that uses data from junior researchers' first years of publication to predict their stable Q values. We tested the deep model and the linear regression model and compared their accuracies. We have obtained reliable results showing that the predicted values estimated with both models are better than the estimated Q values computed with the Q model itself when using only data from the first five years of publication. Lastly, we note that both approaches are robust approaches to deal with the inflation of citation bias. (c) 2021 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据