Intrusion Detection in the Era of Machine Learning: A Critical Survey of Algorithms and Evaluation Practices
Bhavika Nanu, Neelam Duhan · International Journal of Computer Applications · 2025
With the growing prominence and sophistication of cyberattacks, IDS are now indispensable in securing computer networks.Traditional signature-based methods often fail to detect novel threats, prompting the adoption of ML and DL techniques into IDS.This review explores a range of ML algorithms: such as Decision Trees, Random Forest, Support Vector Machines, k-Nearest Neighbors, Naïve Bayes, and Logistic Regression-as well as DL models like Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN).It explains their use in anomaly detection with established datasets like NSL-KDD and UNSW-NB15, and emphasizes importance of data preprocessing, feature selection, and evaluation measures (precision, accuracy, recall, F1-score).The survey emphasizes the strengths as well as constraints of every method, indicating that ensemble & deep learning methods show improved detection accuracy.Finally, it outlines key challenges and proposes future research avenues for developing robust & intelligent IDS solutions.