%A BAI Rujiang,ZHAO Mengmeng,ZHANG Yujie,DONG Kun %T Review on Scientific Literature Mining: Tools and Technologies %0 Journal Article %D 2021 %J Frontiers of Data and Computing %R 10.11871/jfdc.10-1649.2021.06.005 %P 60-80 %V 3 %N 6 %U {http://www.jfdc.cnic.cn/CN/abstract/article_175.shtml} %8 2021-12-20 %X

[Objective] By reviewing on the main tools, system platforms and key technologies of scientific and technological literature mining, this paper presents the future development trend and the reference to the relevant researches.[Methods] By literature review and network survey, this paper summarizes the historical roadmap of scientific and technological literature mining, introduces the main tools, system platforms, and characteristics of scientific and technological literature mining, makes a comparative analysis from the aspects of platform function, data type, and processing ability, and focuses on the key technologies and the state of art of the scientific and technological literature mining. [Results] This paper resolves the problem of "where to mine, what to mine, and how to mine" in scientific and technological literature mining. The data source of scientific and technological literature mining is evolving towards multi-source data fusion and fine-grained knowledge organization. The construction of semantic knowledge map of scientific and technological literature is a hot topic. The graph neural network, pre-trained model, generative adversarial networks are the frontier key technologies. The causal inference methods are becoming the frontier direction. [Conclusions] With the continuous in-depth development of big data and artificial intelligence, scientific and technological literature mining is of a great value in specific application scenarios such as scientific and technological information decision-making with the help of data and technological benefits.