Affiliations: [a] State Grid Shandong Electric Power Research Institute, Jinan, Shandong, China | [b] School of Computer Science and Technology, Shandong University of Finance and Economics, Jinan, Shandong, China
Correspondence:
[*]
Corresponding author: Lida Zou, School of Computer Science and Technology, Shandong University of Finance and Economics, Jinan, Shandong 250014, China. E-mail: [email protected].
Abstract: The data of scientific research is important for decision making of sci-tech journals. At present more and more journals attach great importance to their impacts. As the data expands and artificial intelligence develops, it is feasible to help select topic for journals. In the paper we propose a topic selecting framework based on the research focus, which includes analysis of both research focus and paper topics, model training and topic recommendation. Compared with deep neural network, the proposed method could be more accurate with various periods and time intervals. We also predict different models for different kinds of paper. In this way, the model is more adaptable. The experiments demonstrate that our method is 69.8% better than benchmark algorithm in accuracy rate.
Keywords: Sci-tech journals, research focus, journal impact, prediction, topic selection
