Paper Title
IDENTIFICATION OF CYBER ATTACKS USING BACK PROPAGATION NEURAL NETWORK

Abstract
Cybercrime is spreading rapidly around the globe, and its perpetrators are taking advantage of every conceivable flaw in the way computers are designed and maintained. Ethical hackers focus more of their efforts on identifying vulnerabilities and offering methods for mitigating such vulnerabilities. The community working in the area of cyber security has been expressing an urgent need for the creation of efficient approaches. The majority of the methods that are utilised in IDS systems today are not equipped to cope with the dynamic and complicated nature of cyberattacks that are launched against computer networks. As a result of machine learning's efficacy in addressing various cyber security challenges, the topic of machine learning for cybersecurity has lately taken on a greater level of significance. Techniques from the field of deep learning have been used to address important concerns in the realm of cyber security, including the detection of intrusions, malware, spam, and phishing. Although machine learning cannot fully automate a cyber security system, it does assist to detect cyber security risks more effectively than other software-oriented techniques, and as a result, it relieves some of the pressure that is placed on security analysts. Consequently, effective adaptive approaches, such as the many different techniques of deep learning, have the potential to result in greater detection rates, reduced false alarm rates, and acceptable computing and communication costs. Our primary objective is to ensure that the process of locating assaults is fundamentally distinct from those of these other applications, making it substantially more difficult for the community concerned with intrusion detection to make successful use of machine learning. Cybersecurity is now experiencing new issues as a result of the expansion of cloud computing services, the rise in the number of individuals using online applications, and the alterations to network architecture that link devices that run several operating systems. For this reason, network security components, sensors, and insurance conspiracies need to be built in order to suit the demands and worries of consumers in order to battle emerging threats. In addition, customers' wants and concerns must be taken into consideration while developing these components. In this piece, we will concentrate on countering application layer cyber assaults, which are considered as the most hazardous threats and the most essential test for network and cyber security. [Cyber] attacks on application layers are becoming more common. The majority of this article is dedicated to discussing machine learning as a strategy for coping with routine model usage and identifying potential cyber risks. In order to get examples in the form of Perl Compatible Regular Expressions (PCRE) ordinary expressions, the chart-based division approach and dynamic programming were used. The client sends HTTP queries to a web worker, which is then included into the model. These requests are used to collect information. Keywords - Vulnerabilities, Network security.