Government affair evaluation method, apparatus, and device based on artificial intelligence, and storage medium

Abstract

This invention discloses a government affair evaluation method based on artificial intelligence and relates to the field of government affair text processing. Content correlated to a target department is recognized from superior work report, and a work report of the target department is obtained, to form target text. Public opinion information of network is crawled. The target text and the public opinion information are input into a pre-trained summarization model respectively, to obtain a report summary and a public opinion summary. All public opinion summaries obtained are clustered to obtain public opinion clusters. All public opinion summaries in each public opinion cluster are merged and a summary is regenerated, to obtain a cluster summary. Similarity between the report summary and each cluster summary is calculated, to obtain a government affair evaluation score of the target department. Corresponding devices, equipment and storage media are also provided