Empowering Security: Machine Learning Solutions for Detecting Brute Force Attacks

It is essential to identify brute force attacks, which are a common type of cyber security threat, to protect networks and desktops against unauthorised entry. Previous rule-based methods frequently face difficulties in adjusting to the changing strategies and intricacies of contemporary attacks. This paper presents a machine learning(ML) approach to detect brute-force attacks by utilising predictive modelling and anomaly detection technology. The technique seeks to detect suspicious activity that is symptomatic of brute force attacks by analysing trends in unsuccessful logins, user behaviours, and network traffic. Utilise a range of ML algorithms, such as Logistic Regression (LR), Naive Bayes (NB), Decision Tree (DT), and K-Nearest Neighbour (KNN), to train predictive models on labelled datasets containing past attack data. In addition, investigate feature design strategies, model parameters, and assessment metrics to enhance the effi-ciency of the detection system. By conducting experiments and evaluating real-world datasets, want to prove the efficiency and scalability of the suggested method in effectively detecting brute force attacks while reducing the occurrence of false positives. The findings emphasise the capacity of ML techniques to strengthen cybersecurity measures and reduce the threats presented by brute-force attacks in the current digital environment. The NB, KNN, and LR classifiers have accuracy rates of 87.51%, 99.80%, and 95.55%, respectively. The DT classifier achieves the greatest accuracy rate of 99.96%.