NETWORK SECURITY PREDICTION MODEL USED BY NEURAL NETWORKS

Abstract

 The article discusses four algorithms, namely the SVM algorithm, the fuzzy clustering algorithm, the K-Means clustering algorithm, and the Apriori algorithm. We detail the 4 different steps of network user security and their access control. The article is the development of a reliable network security prediction model. An intrusion detection model built using neural networks has been developed. The intrusion detection model detects anomalies and abuse-based attacks. The intrusion detection model also performs three types of classification tasks. Tasks include classification between the occurrence of an attack or a normal case, classification between different types of attack or a normal case. The intrusion detection model also shows classification accuracy, execution time, and memory usage. The goals of the intrusion detection model are high accuracy, low execution time, and minimal memory usage. An intrusion detection model built using neural networks meets the goals of high accuracy, low execution time, and minimal memory usage.

In today's world, networks are becoming increasingly complex, interconnected, and widely used. Today, network traffic is growing almost exponentially. Networks also become more vulnerable to attack by hackers or anyone with malicious intent to disrupt network systems. Vulnerable networks are at risk of a blow to the economy and the destruction of confidential information. Thus, there is a need to improve network vulnerability detection mechanisms and improve network security prediction. The network security prediction model also aims to reduce memory consumption and improve the detection of different types of attacks in terms of timing and accuracy. In the network security prediction model, the memory consumption was low, and the time spent to detect attacks was also low. Attack detection accuracy is also high. The above methods used to design the model are also easy to design. The above methods are also much more cost-effective since the use of neural networks is free. In addition, calculations are simplified by using this model. Therefore, using a neural network is also an effective way to develop a network security prediction model. Thus, the use of neural networks is recommended for the development of any type of network security prediction model. Future tasks are to develop models that will detect any intrusions even more accurately and quickly.