Research On The Mechanism Of News Recommendation Algorithm For Domestic Platform-based Media

Personalized algorithm recommendation is one of the most concerned technologies in the mobile Internet era.Some news and information companies or media not only produce their own content,but also allow them to absorb many non-platform production content,such as UGC,to distribute content by means of algorithms.This article selects three platforms,namely People' s Account,Sohu News,and Today' s Headline,to study the news recommendation algorithm mechanism of domestic platform media.The People' s Algorithm is contracted to Baidu,referring to the Baidu News recommendation algorithm mechanism.The domestic platform media news recommendation algorithm is divided into three dimensions:core algorithm iteration,news element weight,and algorithm value.The study found that today' s headline pursues "100%accurate portraits",Sohu makes conservative personalization" recommendations,and Baidu is committed to using technology to reproduce realistic logic to become the "smartest" personalized digital tool.Domestic platform-based media imitates domestic and foreign algorithm models in terms of algorithm update iterations.The parameters are roughly similar.In.the future,we need to worry about the same social problems caused by algorithm homogeneity.At the same time,the algorithm still has unexplainable problems.Solve the problem of unexplainability of the algorithm and improve the trust of the algorithm.