Fecha de publicación: 22/10/2021 Fuente: Wipo "BigData" The invention discloses a hot news prediction method based on a medium contribution degree. The method comprises the following steps: 1, extracting a news cluster in a news data set by utilizing a hot news label generation method and system; 2, acquiring related data of the hot public opinion events by utilizing an intelligent information processing technology; and 3, calculating the hot events in different fields by using a contribution value calculation model. According to the method, the defects that an existing manual method is low in efficiency and the accuracy seriously depends on knowledge and experience are overcome, the method is achieved through the big data and semantic analysis technology and a computer algorithm, the speed, efficiency and application scenes are greatly improved, mass data are collected and analyzed through the big data technology, sample data and cases of analysis are greatly expanded, a large number of cases accumulated historically are fully utilized, characteristics of various aspects of user content tendency and public opinion hotspot propagation are mined, the model is more scientific and reasonable, and analysis results are continuously improved and reach certain accuracy.