Research on Computer Network Big Data Security Defense System Based on Support Vector Machine and Deep Learning

Due to the swift progress of big data technology, traditional network security systems that are in use can hardly deal with complex and diverse network attacks. In the current paper, a network security system that is based on big data technology is proposed. By the means of limited mathematical computations, for example, genetic algorithm (GA), support vector machine (SVM) and deep learning (DL) a superior and smarter defense system is created. Above all, the system carries out extensive data collection, that is traffic data, user behavior and system log in real time, with the help of numerous distributed data processing technologies such as Hadoop and Spark. The genetic algorithm was then employed to create the features required for the intrusion detection system to optimize the feature set of the detection systems which in the turn manifolds the detection efficiency. Next, to classify the network traffic, the support vector machine is utilized, and after that, optimization of the hyperplane separation is performed to enhance the accuracy of the anomaly detection. Lastly, the deep learning technology was used to significantly analyze the network traffic and it automatically produced higher levels of features and thereby enabling the machine to extract complex attack behaviors. The experiment result shows that the detection accuracy of our proposed system is 95. 4%, 92.1% and 89.7% when it comes to DDoS, SQL injection and XSS attacks, respectively, while the response time is much better than that of traditional defense systems. This paper's analysis supplies conceptual underpinning and practical protection for the issue of safety within the network of big data, and also has a significant practical impact.