Convolutional Neural Network-Based Anomaly Detection for Cyber Security: A Survey

There are many security risks as computer and communication technologies continue to advance. To reduce these worries, several different kinds Numerous processes with cyber-security enabled have been developed. Anomaly detection is one of these methods. Finding anomalies involves employing a variety of approaches and procedures to identify various patterns that do not adhere to predetermined characteristics of the entire body of data. Deep learning techniques have recently been accepted as an adequate option due to their capacity to take data properties directly from the data. Convolution neural network (CNN) is employed primarily because it can handle input with many dimensions. This study offers a thorough analysis of the use of CNN as a primary tool for anomaly identification. The majority of the known remedies have been compiled and categorized in the literature. Additionally, this research provides a bound together cross framework that mimics complete anomaly detection methods that were previously offered examinations, as per the information source. In-depth assessment from a brought together cross system upgrades this paper’s top to bottom conversation of how CNN is utilized in the writing’s response to the irregularity identification issue. Finally, this article makes a number of recommendations for future study areas that the audience may find useful in their own further activities in this area.