Detailed Analysis of Intrusion Detection using Machine Learning Algorithms
Samriddhi Verma1, Nithyanandam P.2

1Samriddhi Verma*, School of Computer Science and Engineering, Vellore Institute of Technology (VIT University), Chennai, India.
2Dr. Nithyanandam P., School of Computer Science and Engineering, Vellore Institute of Technology (VIT University), Chennai, India. 

Manuscript received on April 30, 2020. | Revised Manuscript received on May 06, 2020. | Manuscript published on May 30, 2020. | PP: 1894-1899 | Volume-9 Issue-1, May 2020. | Retrieval Number: A2127059120/2020©BEIESP | DOI: 10.35940/ijrte.A2127.059120
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: The number of internet users has increased exponentially over the years and so have increased intrusive activities significantly. To detect an intrusion attack in a system connected over a network is one of the most challenging tasks in today’s world. A significant number of techniques have been developed which are based on machine learning approaches to detect these intrusion attacks. Even though these techniques are good, they are not good enough to detect all kinds of attacks. In this paper, the analysis of different machine learning algorithm will be performed on the NSL-KDD dataset with pre-processing steps like One-hot encoding, feature selection and random sampling to use in different machine learning models to find the best performing model to detect these attacks. The attacks are from the datasets are classified into four types of attacks: Probe, DoS, U2R, R2L while the non- attack is the Normal. The dataset is in two parts: KDD-Train and KDD-Test. The dataset is trained and tested to find accuracy and understand the performance of different machine learning algorithms and compare them. The Machine Learning algorithms used are Naive Bayes Classifier, Decision Tree Classifier, Random Forest Classifier, K-Neighbours Classifier, Logistic Regression, SVM Classifier, Voting Classifier. These techniques are compared according to their capability to detect the attacks. This comparison will help to find the algorithm which would work the best to detect different kinds of intrusion attacks. 
Keywords: Intrusion Detection, NSL-KDD, Supervised Learning, One-Hot Encoding, Feature Selection.
Scope of the Article: Machine Learning