Citation Counts Prediction of Statistical Publications Based on Multi-Layer Academic Networks Via Neural Network Model

Citation counts is a crucial factor for evaluating the quality of a research paper. Therefore, it is of vital importance to accurately predict citation counts and to explore the mechanism underlying citations. In this paper, we mainly focus on the prediction of citation counts in the field of statistics. We collect 55,024 academic papers published in 43 statistics journals from 2001 to 2018. Our main contributions are as follows. First, we invest considerable effort in collecting and cleaning a high-quality dataset. Second, multi-layer networks are constructed from different perspectives, including journal network, author citation network, co-citation network, co-authorship network, and keyword co-occurrence network. In addition, we extract 77 factors for citation counts prediction, including 22 traditional factors and 55 network-related factors. Third, to address the issues of zero-inflated and over-dispersed citation counts, a neural network model is designed to achieve high prediction accuracy. Furthermore, we adopt a leave-one-feature-out approach to investigate the importance of factors. The results show that our neural network model outperforms the other methods. This study provides a useful guide for researchers to predict citation counts and can be easily extended to other research fields.