Intrusion Detection System monitors the network resources in an effort to find unsuitable network activity. Attackers constantly create new vulnerabilities and attack strategies intended to weaken the defense. Obtaining user credentials that give access to the network and data is the goal of several attacks. For network security, a Network Intrusion Detection System is crucial since it enables to react to hostile activity. With the growth of the Internet, attackers, spammers and criminals have become a greater menace. Due to the frequency of such dangers, intrusion detection systems (IDS) have emerged as a crucial component of computer networks. Most of the previous IDS have focused on the classification of attacks as normal or intrusion. In this paper, a novel approach of creating intrusion detection system is created using different algorithms of machine learning. The proposed system can classify the type of attacks. We have used three types of classifier models SVM, SGD, and Adaptive Boost with a different set of features. The method is examined using knowledge discovery on the KDDcup99 dataset. The findings demonstrate that the suggested system has a low false alarm rate and a high accuracy of attack detection.