A feature aggregation-based method for identifying rumors in social networks

Abstract

The invention discloses a method for identifying rumors in social networks based on feature aggregation. By designing acceptable time series propagation mode features and time series text features of a deep neural network, and using feature aggregation technology, a rumor detection model is constructed to carry out the finalization of rumors. detection and early detection. The present invention solves the problem that the propagation mode feature of social network event propagation is difficult to input into a machine learning model. The propagation mode feature does not depend on feature engineering and domain knowledge, is a comprehensive reflection of the influence of various factors in the actual propagation process, and can be effectively applied to Different rumor identification scenarios; avoid the defect that the quality of feature data is degraded due to the huge difference in the number of messages contained in different samples, and solve the problem that a single model in traditional machine learning methods is difficult to deal with heterogeneous information. Compared with previous rumor identification methods Significant accuracy improvement.

Description

technical field [0001] The invention belongs to the field of artificial intelligence, and particularly relates to a feature modeling and detection method of social network rumor information. Background technique [0002] With the development of social networks, the amount of information has grown dramatically. However, the quality of information cannot be guaranteed, and false information represented by rumors has penetrated into almost every corner of the social network. Therefore, how to realize automatic information credibility assessment and predict the authenticity of social media information has high practical significance. [0003] Identification of unknown rumors is one of the urgent requirements for information credibility assessment and information content security. Social psychology defines rumors as news whose authenticity is unverified or intentionally false. The spread of rumors is detrimental to people's lives and social stability, and may cause unexpected l...

AI Technical Summary

Problems solved by technology

Feature extraction is a process of obtaining sample distribution characteristics by using statistics and other methods. There are differences in sensitive feature types in different scenarios. Therefore, traditional machine learning methods rely heavily on feature engineering and are not suitable for complex and changeable social networks.

Secondly, the feature extraction process of traditional methods also requires rich domain knowledge, including the attributes of rumors, the current network environment structure, etc., but due to human interference, it will inevitably cause deviations.

Images