Abstract: 【Purpose/significance】Evaluation science is an important part of the philosophy and social science system. The construc⁃tion of its method system needs to rely on the support of modern technology. An evaluation science method system can greatly improve the efficiency of evaluation activities. The development of deep learning technology has brought unprecedented opportunities for the construction of the evaluation scientific method system.【Method/process】Through bibliometrics,text analysis and other methods,this paper uses ITGInsight to sort out the development status of deep learning and its application in library & information science and evaluation science,and conducts research on methods of evaluation science based on deep learning.【Result/conclusion】Eventually,
the current research situation in this field,as well as the main influence and potential development opportunities brought by deep learn⁃ing technology to the construction of evaluation scientific method system,were finally analyzed. Finally,a method system frame based on deep learning technology was creatively constructed.【Innovation/limitation】This paper has a framework-level guiding role for the research on methods of evaluation science based on deep learning, but has not discussed and empirically researched the specific method process,which will be the content that the related research needs to pay attention to next.