Network intrusion detection is a key element in securing computer networks against unauthorised access and cyber-attacks. Traditional network intrusion detection systems (NIDS) face challenges in identifying unknown sophisticated threats due to their reliance on signature-based detection. In contrast, anomaly-based automated detection shows significant superiority. Machine learning approaches show promise in identifying unknown malicious attacks. This research presents an innovative ensemble-based machine learning technique for network intrusion detection, using the CICIDS 2017 dataset. Leveraging technological progress, and using several ensemble strategies, such as Random Forest, Adaboost, Bagging, Bernoulli Naive Bayes, Gaussian Naive Bayes, KNeighbors classifier, Decision Tree, and Multinomial Naive Bayes, we evaluate the performance of the proposed approach. Correlation analysis is employed to select relevant features for network intrusion detection. Our research, utilising various ensemble methods, reveals that the Random Forest technique consistently outperforms existing methods, achieving accuracy and False Positive Rate (FPR) levels typically exceeding 99.9%. This innovative strategy, incorporating robust evaluation metrics, emerges as a valuable tool for fortifying computer systems and networks against emerging cyber threats.