Sensitivity analysis of the online fake news identification

This paper conducts sensitivity analysis on rumor recognition methods, mainly to exclude the potential impact of content themes, user types, and datasets on rumor recognition. Firstly, the content and themes of rumor information vary, and can be divided into political rumors, economic rumors, and so on. It is necessary to analyze whether the network structure parameters of each type of rumor and fact are different; Secondly, most of the factual information used in this article is created from official accounts, and it is necessary to analyze whether the network structure parameters of official and unofficial factual information and rumor information are different.