Abstract: 【Purpose/significance】Online public opinion is one of the emerging forms of collective intelligence.It is of great signifi‐
cance to explore the public cognitive law based on the online public opinion from different platforms, which can provide theoretical
support for the guidance of public opinion【. Method/process】This study takes Sina Weibo and Tik Tok as data sources.We extract enti‐ties and relationships based on the ELECTRA and REDP respectively, and construct the knowledge graphs of online public opinion for different platforms.We compare and analyze the differences of the knowledge graphs of different platforms【. Result/conclusion】The re‐sults show that the public opinion of microblog and short video contains the same mainstream content, but also has different derivative content.In terms of information content, the public opinion of Sina Weibo platform has richer connotation and wider extension than the Tik Tok platform.In the contribution and acquisition of knowledge, users of Sina Weibo platform are more inclined to participate in the discussion of public opinion and actively contribute knowledge than the other.In terms of knowledge graph construction, the joint prob‐ability of knowledge entities in public opinion of short video platform is higher, which is easier to form knowledge triples, and public opinion knowledge is more focused【. Innovation/limitation】This study constructs and compares the public opinion knowledge graphs of different platforms.In the future research, we will consider the process and causes of the differences of the knowledge graphs of dif‐ferent platforms。