Multi-Source Data Fusion Study in Scientometrics

Hai-Yun Xu; Chao Wang; Hong-shen Pang; Li-jie Ru; Shu Fang

Hai-Yun Xu Chengdu Library of Chinese Academy of Sciences, Chengdu, Sichuan - Institute of Scientific and Technical Information of China
Chao Wang Chengdu Library of Chinese Academy of Sciences, Chengdu - University of Chinese Academy of Science
Hong-shen Pang Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences
Li-jie Ru Chengdu Library of Chinese Academy of Sciences, Chengdu - University of Chinese Academy of Science
Shu Fang Chengdu Library of Chinese Academy of Sciences

Abstract

This review provides an introduction to MSDF, and discusses the status quo of the methods and applications of MSDF in scientometrics study. Currently, one of the most widely used methods of MSDF in scientometrics is the linear mode with a simple and random fusing process. In this paper, we assume that the improvement of MSDF methods requires a strong mathematical foundation, and breakthrough of MSDF in future may come from advanced fields in data fusion research and their applications, such as sensors, the automation, etc. Based on this assumption, this review have investigated the main thoughts of MSDF used in those fields and proposed that MSDF could be divided into the fusion of data types and fusion of data relations. Furthermore, the fusion of data relations could be divided into the cross-integration of multi-mode data and matrix fusion of multi-relational data. This paper studied the methods and technological process of MSDF applicable to information analysis, especially in the competitive intelligence of scientific and technological area.