Bio-syncretic Rehabilitation Mechanism Theory And Application

At present,various attacks against computers emerge in endlessly,and anomaly detection can not only detect unknown attacks,but also can be used to detect internal threats.Compared with traditional machine learning models,deep learning models are more capable of digging out the association relationships between data features.Therefore,this paper adopts the idea of deep learning and proposes an effective network traffic anomaly detection method(GMA-MTL),which performs data preprocessing according to the characteristics of network traffic data and establishes a multi-task neural network classifier to detect network traffic anomalies.First of all,this article uses the ADASYN adaptive sampling algorithm to deal with the imbalance problem in the network traffic data set.In network traffic data,due to the differences in the attack behavior of each abnormal category,there will be more obvious data imbalances when capturing abnormal traffic data,which will bring greater misleading to the experimental results.This article analyzes the pros and cons of commonly used data balancing algorithms,and finally selects the ADASYN algorithm for data balancing.Secondly,this paper proposes a hybrid feature selection algorithm combining embedded and filtered to solve the problem of high-dimensional features in network traffic data.The first step is to use the embedded feature selection algorithm to find the more important feature combinations,and the second step is to use the filtering feature selection algorithm for the remaining feature sets to filter out redundant or unrelated features that are not related to the classification target.Obtaining the optimal feature subset through the hybrid feature selection algorithm can improve the efficiency of the anomaly detection method and improve the model prediction results.Finally,this paper builds a multi-task learning classifier(MTL)and trains and predicts the anomaly detection method proposed in this paper on the UNSW-NB15 public network traffic data set.Use relevant evaluation indicators to evaluate the prediction results,and then use comparative experiments to verify the effectiveness of the GMA-MTL network anomaly detection method proposed in this paper.