Evaluation of Machine Learning Algorithms Performance in Comparison to Cloud Privacy Legislation
Sourabh Patil, Vishal Bhatnagar, Shyla Singh · 2023
An Intrusion Detection System (IDS) is a security tool that checks network data for any indication of malicious activity and notifies the network administrator. The proposed system puts into practice an anomaly-based IDS using a variety of supervised machine approaches. Anomaly and normal, the binary classification employed by the IDS model, are used to categorize, or identify anomalous network packets (0). The model evaluation is compiled using metrics for machine learning evaluation. The anomaly-based IDS model outperforms signature-based IDS in terms of its high computational cost, confined coverage, lack of adaptability, and high false positive rate. The dataset is utilized to train our IDS Model and it is trained and evaluated using various classifiers. Notably, the Random Forest Classifier achieves an exceptional accuracy rate of 99.92%, surpassing all other classifiers, thereby aligning with the cloud privacy legislation's objective of ensuring data confidentiality and integrity.