The usage of the Internet, the amount of network traffic volume is rapidly raising high. Preventing such huge, sensitive data from security attacks is equally important. Monitoring network traffic may indicate a possible intrusion in the network and therefore anomaly detection is important to detect and prevent such attacks. One of the approaches to anomaly detection is based on machine learning classification techniques. Here we apply seven different machine learning techniques: K-Means, K-Nearest Neighbours (KNN), Fuzzy C-Means (FCM), Support Vector Machine (SVM), Naïve-Bayes (NB), Radial Basis Function (RBF) and Ensemble method (Weighted average method) comprising of K-Nearest Neighbours and Naïve-Bayes (NB) on NSL-KDD dataset and evaluate the performance of these techniques. We also deduced how the change in the training size can affect the training time and the accuracy of each algorithm