Research on abnormal detection method of network traffic based on unbalanced data
To effectively address the impact of imbalanced data and ensure real-time detection of anomalies in large-scale network traffic data,a network traffic anomaly detection method based on imbalanced data is proposed.The network traffic data containing imbalanced data is balanced by optimizing the SMOTE(Synthetic Minority Oversampling Technique)algorithm.The obtained dataset is then feature extracted using kernel principal component analysis and input into a convolutional neural network.Further achieve deep feature extraction of network traffic data through convolution and pooling processes,classify network traffic features based on the Softmax classification layer,and use the trained convolutional neural network prediction model to achieve network traffic anomaly detection of imbalanced data.Through experimental verification,this method demonstrates good efficiency and stability.When the number of iterations is 40,achieve the best unbalanced data processing result,and ultimately be able to accurately identify abnormal data.
万方数据
维普
