• Aiswaryaa Velumani
• Harini Chokkanathan
• S.Vijayalakshmi

Monitoring network traffic for intrusions has always been a difficult undertaking. Intrusion detection systems (IDSs) face significant hurdles because of the variety of network threats. Traditional attack recognition techniques frequently use data mining to discover abnormalities, which has the drawbacks of a high false alarm rate (FAR), a low recognition accuracy (ACC), and a limited capacity for generalization. As Machine learning advancements are opening the door for improving intrusion detection systems, we propose an Intrusion Detection System that uses Feature selection to enhance the comprehensive capabilities of IDS and boost network security. Our system uses a feature selection approach that integrates Principal Component Analysis, a well-liked unsupervised learning method for reducing the dimensionality of data, and Recursive Feature Elimination, which fits a model and eliminates the weakest feature. These two methods make learning more convenient and effective. The KDD and UNSW datasets are input into the system to train and evaluate the Deep Neural Network's capabilities and requirements, respectively. This improves IDS' formalism and clarity for the Internet of Things and Cloud Computing, which are booming in the 5G-ready technological industry right now.

IDS, RFE, PCA, DNN

"A Novel and efficient Intrusion detection strategy using effective feature selection", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 5, page no.h379-h385, May-2023, Available :http://www.jetir.org/papers/JETIR2305744.pdf

"A Novel and efficient Intrusion detection strategy using effective feature selection", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 5, page no. pph379-h385, May-2023, Available at : http://www.jetir.org/papers/JETIR2305744.pdf