Chinese spam recognition method based on chaos particle swarm optimization CNN

Abstract

The invention protects a Chinese spam recognition method based on a chaos particle swarm optimization CNN (convolutional neural network). According to the method, first, a word segmentation device isused to perform word segmentation, stop word removal and other preprocessing on a Chinese spam dataset; second, a Word2vec model is adopted to acquire word vectors, and text vectors of Chinese spam are obtained by solving the sum and the mean value of the word vectors; third, the chaos thought is introduced into a particle swarm optimization algorithm to train network parameters of the CNN; fourth, a Chinese spam classification model is established based on the chaos particle swarm optimization CNN; and last, a test set is adopted to realize spam classification through the established model, and the classification correct rate is calculated. Through the method, quick convergence can be realized through the model established according to optimization parameters of the chaos particle swarm optimization algorithm, good robustness and stability are achieved, and meanwhile the classification recognition rate of Chinese spam can be increased.

Description

technical field [0001] The invention belongs to the technical field of Chinese spam classification, in particular to a method for identifying Chinese spam based on a chaotic particle swarm optimized CNN network. Background technique [0002] The proliferation of Chinese spam has seriously endangered the development of my country's Internet information technology. Therefore, establishing an effective Chinese spam classification and identification model to distinguish spam from normal mail will help reduce the waste of network resources and storage space and maintain the Internet information environment. stable development. [0003] Commonly used spam classification methods include neural network (Neural network, NN), Bayesian theory (Bayes), decision tree and support vector machines (Support vector machines, SVM) and so on. Due to the large number and variety of Chinese spam, and the diversity of spam content forms, traditional machine learning methods are limited when proces...

AI Technical Summary

Problems solved by technology

Due to the large number and variety of Chinese spam, and the diversity of spam content forms, traditional machine learning methods are limited when processing a large number of data samples, and it is difficult to establish an efficient classifier model. At the same time, in practice Also hampered by application issues

Images