A strategy for co-authorship recommendation: analysis using scientific data repositories

Abstract

In a co-authorship network papers written together represent the edges, and the authors represent the nodes. By using the concepts of social network analysis, it is possible to better understand the relationship between these nodes. The following question arises: “How does the evolution of the network occur over time?”. To answer this question, it is necessary to understand how two nodes interact with one another, that is, what factors are essential for a new connection to be created. The purpose of this paper is to predict connections in co-authorship networks formed by doctors with resumes registered in the Lattes Platform in the area of Information Sciences. To this end, the following steps are performed: initially the data is extracted, later the co-authorship networks are created, then the attributes to be used are defined and calculated, finally the prediction is performed. Currently, the Lattes Platform has 6.1 million resumes from researchers and represents one of the most relevant and recognized scientific repositories worldwide. Through this study, it is possible to understand which attributes of the nodes make them closer to each other, and therefore have a greater chance of creating a connection between them in the future. This work is extremely relevant because it uses a data set that has been little used in previous studies. Through the results it will be possible to establish the evolution of the network of scientific collaborations of researchers at national level, thus helping the development agencies in the selection of future outstanding researchers.