Network water army behavior detection method and system based on mixed Dirichlet process

Abstract

The invention relates to a network water army behavior detection method and a network water army behavior detection system based on a mixed Dirichlet process. The network water army behavior detection method includes: collecting original data comprising user behavior features and content features, performing quantization representation on the original data of each dimensionality so as to form history behavior vectors of a user, and forming a user data set to be clustered; clustering the history behavior vectors in the user data set to be clustered so as to obtain at least one classification user behavior set; converting data in the at least one classification user behavior set, and merging data with the same user identification in the data after being converted so as to obtain a sequence database; performing sequence pattern excavation on the sequence database through a pattern excavation module so as to obtain at least one affair sequence pattern corresponding to each classification user behavior set; judging out the classification user behavior set which is a water army username by comparing each affair sequence pattern through a water army judging module. The network water army behavior detection method and the network water army behavior detection system based on the mixed Dirichlet process can easily recognize which category belongs to the water army username.

Description

technical field [0001] The invention relates to the field of network security, in particular to a method and system for detecting network troll behavior based on a mixed Dirichlet process. Background technique [0002] The popularity and rapid development of social networks not only brought fundamental changes in information communication technology, but also changed people's production and lifestyle, interpersonal communication and way of thinking step by step, affecting all areas of social life and the survival of human beings. It has had a wide-ranging and profound impact on development. At present, social networks include all forms of network services centered on human social interaction. From a historical perspective, it will be a key force that will push the Internet closer to the real world. [0003] With the rapid development of social networks, the security problems it brings are becoming more and more prominent, among which the security problems originating from t...

AI Technical Summary

Problems solved by technology

The existing "network trolls" generally use algorithms such as decision trees, KNN, and neural networks. During the design process of these algorithms, it is necessary to set the values ​​of key parameters based on historical experience to reflect the impact of every aspect of user behavior. The degree of influence of the judgment results is different. This method has a lot of subjectivity and seriously affects the judgment accuracy of the algorithm.

Images