Abstract: 【Purpose/significance】Network public opinion early warning.as a‘barometer’and‘reminder’to reflect social public opin⁃
ion.is helpful for government departments to manage society scientifically through announcement.communication.emotional comfort and educational activities【. Method/process】Based on Python data crawler technology.38 days during the epidemic in Equatoria Prov⁃ince of Congo (Kinshasa) were divided into 38 time points.and the SVM model was further constructed.and the SVM model was trained with Matlab.The training samples were from May 18 to June 18 and the test samples were from June 19 to June 25【. Result/conclusion】
Through empirical research.the situation in which the degree of crisis is greater than or less than 0.5 is reasonable and effective.and the early warning model is practical.which has great value for the government.society and media to deal with the crisis【. Innovation/lim⁃itation】However.since this article is only for the Ebola epidemic.the amount of data obtained from Twitter is limited.so there are cer⁃tain research limitations.In the future.we will try to extract more data from multiple sources based on more topics to conduct a more sys⁃tematic and comprehensive research on the online public opinion crisis early warning mechanism.