摘要
Abstract
In large-scale and highly complex network environments,data transmission is subjected to interference such as noise and multipath effects.These interference factors lead to a decrease in the quality of network transmission data and thus reduce detection effect.In view of the above,a deep learning algorithm based intelligent detection method for network transmission data is proposed.Firstly,a network transmission data view is constructed based on Gaussian noise,masks,and Dropout to enhance the network transmission data.Then,the enhanced network transmission data is input into a deep variational autoencoder to perform encoding and decoding conversion,so as to reduce the dimensionality of network transmission data with high complexity.Finally,the dimensionality reduction results of network transmission data are set as the inputs of the long short-term memory(LSTM)network.By learning the correlation of time series of network transmission data,the network transmission data can be detected intelligently.The experimental results show that the proposed method can accurately detect abnormal network transmission data such as out-of-order data,missing data,and attack data in the network transmission.关键词
深度学习算法/网络传输数据/智能检测/视图构建/自编码器/解码转换Key words
deep learning algorithm/network transmission data/intelligent detection/view construction/autoencoder/decoding conversion分类
电子信息工程
