A Study on Online Detection of micro-blog Rumors Based on Naive Bayes Algorithm

The paper puts forward a rumor detection model based on Naive Bayes and NLP according to the data volume of detecting micro-blog rumors. Firstly, the model processes micro-blog text through NLP methods, including Chinese particles, stop words, and word vectors. Next, the processed text is input into the algorithm model of machine learning based on Naive Bayes. Eventually, the model verifies the authenticity of processed micro-blog information. Compared with manual rumor detection, this detection model reduces processing time for nearly 10h. Also, its accuracy of identifying micro-blog rumors increases by 15% than that of micro-blog rumor detection algorithms based on the decision tree, showing it is more applicable to current micro-blog platforms.