Visual Modeling of Keyword Dimension Reduction in Double First-Class University Funds Based on t-SNE Algorithm

Abstract

[Purpose/Significance] The National Natural Science Foundation's project funding is an important indicator of scientific research capabilities. Analysis of the data of the establishment of the research funds of double first-class universities is helpful to provide strategic support for university construction. [Purpose/Significance] This article studies the keyword data of the National Natural Science Foundation of China from 1998 to 2017. At first we preprocess double first-class universities' data, and then use the t-SNE algorithm in MATLAB to reduce the dimension of the data and visualize the results. This paper models from the time dimension and the unit-dependent dimension, and studies the keyword distribution of double first-class universities' projects in the past 20 years. [Results/Conclusions] The method in this paper is more intuitive than the traditional method based on structured analysis and provides a reference for the formulation of Chinese universities' construction strategies. In addition, other scholars can further model and program based on this research for such purposes as interactive visual modeling and fast and positioning of massive project data to improve scientific research efficiency.

Keywords

• research fund|data mining|scientific and technological intelligence analysis|network visualization|t-sne