Analysis of Machine Learning based IDS for Detection of Anomaly Profiles

Today, cyber attacks on computer networks pose a serious risk to anyone navigating the web. One type of security instrument used to identify potential intrusions into a Network or Host is an Intrusion Detection System (IDS). According to studies, high detection rates and minimal false positives in intrusion detection can be attained by employing machine learning approaches. This paper objective is to evaluate various Machine Learning (ML) algorithms utilized for anomaly profile detection in IDS. The purpose of this research is to examine current studies in IDS that took a Machine Learning (ML) approach, focusing on datasets, ML methods, and metrics. In order to guarantee that the models you're building are fit for IDS use, it's crucial that you use the right datasets. It's also worth noting that the structure of the dataset can have an effect on how well the ML algorithm performs. As a result, the structure of the dataset matters when deciding on an ML algorithm. Thereafter, the metric will offer a quantifiable assessment of ML algorithms' performance on the dataset in question