The capacity to swiftly monitor network data, spot any infiltration patterns and half any harmful impacts of aberrant entry that could ruin the network as known as network Intrusion detection (IDS). The work focuses on detecting and classifying multiclass attacks using a current and imbalanced dataset, presenting a good challenge for improved security. We applied the Random Undersampling technique to the CSE-CIC-IDS2018 dataset in order to address the imbalanced dataset. After resampling this study examines the effectiveness of machine learning classifiers: Random Forest, Logistic regression, K-nearest neighbor and Decision tree based on accuracy, F1-score, precision, and recall, respectively. The created models are put to the test as multiclass classifiers against 14 different kinds off attacks and benign. Out of that random forest gives the 99.98% of accuracy.