Prof. Imran Ahmad
Spark, Auckland, New Zealand

Title: Enhancing Network Security: Machine Learning-Driven Anomaly Detection Using WEKA

Abstract: In the era of escalating cyber threats, ensuring robust network security has become a critical challenge for organizations. Anomaly detection plays a pivotal role in identifying suspicious activities and potential security breaches within networks. This study explores the application of machine learning techniques, particularly ensemble methods, for anomaly detection using the WEKA (Waikato Environment for Knowledge Analysis) platform. The research aims to enhance the accuracy, scalability, and resilience of anomaly detection systems by leveraging WEKA’s comprehensive suite of machine learning tools. The proposed methodology involves data preprocessing, feature extraction, and the application of ensemble techniques such as Bagging, AdaBoostM1, Stacking, Random Forest, and Vote. Empirical analysis demonstrates the effectiveness of these methods in detecting network anomalies, with Bagging and Random Forest emerging as the top-performing algorithms, achieving high correct classification rates and low error metrics. The study evaluates performance using metrics such as Kappa statistics, mean absolute error, and area under the ROC curve, providing a robust framework for assessing detection accuracy. The findings offer valuable insights for cybersecurity professionals, system administrators, and network security practitioners, enabling the implementation of more effective anomaly detection systems to safeguard critical data assets and network infrastructure. By combining the power of ensemble learning with WEKA’s versatile tools, this research advances the field of network security, offering a scalable and efficient solution for real-time anomaly detection in dynamic network environments.

Bio: To be announced soon