Detection Model of News Distortion Based on Chinese-German Bilingual Knowledge Graphs

To reduce the repeated misinterpretation of China’s news events in other countries’ media, the author constructed a detection model of news distortion based on cross-linguistic knowledge graphs. The research object is news on the well-known portal websites in China and Germany. With the help of the affinity propagation based on bilingual knowledge graphs and the bilingual sentiment dictionaries, the author proposes an automatic discovery mechanism of news distortion in the news transmission across languages. Compared with other traditional methods, the BKG-AP algorithm has the best performance in news clustering, and its F1 score is up to 85.2%. The experiments show that the detection model of news distortion based on Chinese-German bilingual knowledge graphs can discover distorted news more quickly, comprehensively and accurately. Therefore, it can be an effective tool for China to know the public opinion abroad that departs from facts in time and take measures to clarify misunderstandings.