Fake News Detection Model Based on Boundary Point Detection Variable Density Clustering

Aiming at the accuracy problem of false news detection in social media, a new detection model is proposed by the boundary point variable density clustering method. The model firstly constructs the feature portrayal index based on the vector space, uses the weight function of the feature words to portray the text data set, extracts the clustering boundary points by combining the distribution characteristics of the data set, and realizes the conversion of the text feature words to the numerical values, so as to construct the false news detection method based on the boundary point detection variable density clustering algorithm. At the same time, based on numerical analysis to compare and analyze the performance status of the method and other detection methods, found that the method has better accuracy. And through in-depth analysis of the relationship between the accuracy of the algorithm and the key parameters such as the nearest neighbor value, the number of texts, word dimension, etc., it is found that: in the case of small number of texts and word dimension, the smaller the nearest neighbor value corresponds to the lower accuracy; on the contrary, the larger the nearest neighbor value, the higher the corresponding accuracy.