Review Paper
Intrusion Detection Classification and Detection by Machine Learning and Deep Learning Approaches: Review
DoS attacks that disrupt important, intelligent services like healthcare can also result in human death due to the disruption of routine services. devices (for example, intelligent refrigerators, smart televisions, and air conditioners) are easily attacked by attackers who exploit their weaknesses to launch denial-of-service attacks. As a result, one of the primary concerns for researchers all over the world is the protection of these devices. Globally, intrusion detection is being studied to fix this matter. IDS are classified into three categories based on their detection capabilities: Depending on a signature, a specification, or an anomaly. Whenever a device or network connections analyses an attack against a signature contained in the inner IDS database, an attack is identified by IDSs. If a device or network operation matches one of the saved signatures/patterns, a warning will be generated. This method is extremely reliable and effective at recognizing identified risks, and its process is simple to comprehend. However, this technique is ineffective in classifying new attacks and discrepancies between current These sorts of assaults do not have a meaningful signature to identify them If the divergence from a specified behaviour profile exceeds a predefined threshold, an anomaly-based intrusion detection system (IDS) issues an alert. Classifying intrusions does not seem to follow a typical pattern, and understanding the whole spectrum of normal activity is not an easy task. Emerging threats may be identified using this method. As a consequence, there is a significant rate of false positives with this method. Routing tables, protocols, and nodes, for example, are all part of the specification-based approach since they are all defined by a set of rules and criteria. It is possible to identify intrusions when network behaviour deviates from standards' specifications, using specification-based techniques. Therefore, specification-based detection is used for the same goal as anomaly detection: to separate aberrant behaviour from normal behaviour.
Published by: Ayush Gautam, Shivani Rana
Author: Ayush Gautam
Paper ID: V8I5-1147
Paper Status: published
Published: September 14, 2022
Full Details