Method and system for network navy account number identification based on particle swarm optimization

Abstract

The invention relates to a method for network navy account number identification based on particle swarm optimization. The method specifically comprises the following steps that step 1, user information of beta users is collected, lambda correlation indexes are extracted from each piece of user information, and therefore beta index vectors are obtained; step 2, weight allocation is conducted on the lambda indexes according to the requirements so that weight vectors can be obtained; step 3, the beta index vectors are ranked according to the size of each index in the index vectors so that lambda sequences can be obtained; step 4, each user is selected to serve as a particle, the particle swarm optimization is applied, and the particles satisfying the requirements are stored in a navy user list; step 5, all suspicious navy users in the navy user list are selected and all neighbors satisfying the requirements are added to the navy user list. At the real-time aspect, the method for network navy account number identification based on the particle swarm optimization is more suitable.

Description

technical field [0001] The invention relates to a method and system for identifying network navy account numbers based on particle swarm algorithm. Background technique [0002] The increasing amount of information on the Internet has brought great challenges to Internet supervision. A large number of Internet promoters are flooding the Internet environment. Incidents have occurred repeatedly, demonstrating the black power behind them-the cyber trolls. The Internet trolls use the Internet environment to create public topics, thereby guiding public opinion, making the influence of certain individuals and units soar, and seeking illegal benefits from it. There are two modes of operation of the Internet trolls, one is to build momentum, and the other is to advertise. Among them, building momentum is to express specific opinions on a specific topic. The number of words for the purpose of building momentum is mostly between 20 and 30 characters, the arrangement is relatively ne...

AI Technical Summary

Problems solved by technology

However, in this process, the selection of the training set has great uncertainty, and the proportion of water army posts in the training set seriously affects the judgment accuracy of the classifier; moreover, water army posts target different topic objects in different time periods , in order to identify troll posts about a sensitive topic within a certain period of time, it is necessary to train the classifier for this specific topic. It can be seen that this method has a lot of work; Posts are reviewed and categorized, seriously affecting the user experience of the forum

Aiming at the problem that the semantic analysis method has a heavy workload and affects the user experience

Images