A Comprehensive Empirical Evaluation of Shallow Machine Learning Paradigms and Dense Neural Topologies for Robust Network Intrusion Detection Architecture

Author(s)	Mr. Aravind Chagantipati
Country	India
Abstract	The optimization of network security controls within massive enterprise cloud backbones necessitates highthroughput, automated inline classification frameworks capable of identifying complex multi-stage attack matrices. Traditional firewalls and rule-based inspection mechanisms fail entirely against zero-day vulnerabilities and morphing payloads. This study conducts an extensive empirical benchmarking analysis that compares traditional shallow machine learning architectures against multi-layered dense deep network configurations. The evaluation utilizes three globally acknowledged benchmark threat repositories: NSL-KDD, CICIDS, and UNSW-NB15. Quantitative analytics confirm that the recurrent sequence tracking capabilities of Long Short-Term Memory (LSTM) cells produce the absolute highest classification efficiency, yielding a peak accuracy of 97.5% and an F1-Score of 97.0%, significantly exceeding classical classifiers.
Keywords	Intrusion Detection Systems (IDS), Deep Learning Topologies, Tabular Classifiers, Long Short-Term Memory, Performance Metrics.
Field	Computer > Network / Security
Published In	Volume 8, Issue 4, July-August 2026
Published On	2026-07-13

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research