摘要：

Abstract:

【Purpose/significance】Government information resources are an important part of our country's information resources. By

analyzing the active push service mode of government information resources, we can alleviate the problem of information overload and

idle resources, and improve the utilization rate of government information resources.【Method/process】Based on the whole process of

government information resources active push, the design idea of "three horizontal and two vertical" is adopted to innovate the mode of

government information resources active push. With LDA model and association rule mining as the main technology, subject extrac⁃

tion and clustering of government information text are carried out, user interest is mined, and information subject and related network

is built. Finally, information and user matching are achieved through similarity calculation to achieve active push.【Result/conclusion】

Through the hybrid push of hot spot information, personal information and related information, we can help the government to find out

as well as make up the missing, private customization and related prediction in information service, and improve the intelligibility, us⁃

ability and understandability of government information.【Innovation/limitation】In this paper, it is innovative to apply LDA model and

association rules to push government information resources at the same time. However, because of the particularity of the context of

government text, semantic mining directly using LDA model can not achieve the best results.