Research On The Language Characteristics Of Internet Rumors

Nowadays,the rapid development of social networks has gradually become a hotbed for the spread and spread of rumors.In this environment,the speed and spread of rumors cannot be underestimated.Internet rumors will not only interfere with people’s daily lives,damage the credibility and interests of certain enterprises,units,etc.,but may also cause large-scale panic and commotion,and have a great impact on national security and social stability.Among the Internet rumors,rumors of health account for a large proportion,which mainly focus on the themes of food safety,the pros and cons of living habits,the prevention and cure of diseases,and are closely related to people’s health.Therefore,this article selects 100 health rumors as the main research object,and combines 100 authoritative and credible similar popular science articles for comparison,explores its characteristics,and proposes for the automatic identification,screening,and manual identification of Internet rumors.Specifically,this article compares the two types of articles from the title and the body.As for the article title,this article summarizes the tendency of the rumors article title and the popular science article title in terms of punctuation,part of speech,sentence patterns,and sentence categories.For the content of the article,this article compares the specific performance of the length and word richness of the rumor article and the popular article,the classification summarizes the distribution characteristics of the two types of article topics,analyzes the specific characteristics of the top thirty high-frequency words in different word categories,The TF-IDF algorithm compares the top 20 keywords of the two types of articles,the Cosine similarity calculation method is used to compare the text similarity between the rumor article and the popular science article and between each other,the text emotion analysis is used to to compare the composition of positive and derogatory and neutral sentences in the two types of articles.After a detailed analysis,based on the previous conclusions,the author puts forward suggestions for identifying and screening rumors of health,including human judgment and computer automatic screening should focus on the content.In addition,the author also briefly introduced the advantages of popular health science articles.The last part is the summary,which summarizes the research results of this paper,puts forward the further improvement of this paper and the prospect of future research.