An Analytical Assessment And Examination Of Supervised Tree Based Classifier In Intrusion Detection System

Network attack worldwide is now a major concern especially with the security threats that is attached to data loose;thus,making the field of network security an interesting area of research in the field of Computer Science.To deal with situations such as data harked and subsequent illegal intrusion into systems,several methods have been put into place by network securities to help in the prevention and eradication of these threats.Among the many methods is a method called the Machine Learning Technique(MLT).This method has been the proposed method used in this research since indeed it is one of the most reliable methods for determining network intrusion detection.The introduction of internet age and the numerous increased in the resources available to many users,the high risk of vulnerability has increase by the day and is becoming a challenging issue for computer users.Therefore in this paper quest to compare and evaluate intrusion performance,decide to introduce different tree based classification algorithms which has the capacity to classify network procedures in intrusion detection systems under a supervised approach,one of the three types of Machine Learning Classification Procedures.This Analysis were carried out using the following tree based classifier,C4.5,ADTree,RandomTree,RepTree,and RandomForest.A supervised filter known as Discretization will was applied at the pre-processing stage.Discretization in data mining is the process that transforms quantity data into quality data.This paper employs the use of two different feature selection approaches which are:Consistency Subset Evaluation and Correlation Feature Subset Evaluation for the purpose of dimensionality reduction of the dataset.To show the performance of the proposed evaluation method,the experiment is expected to be performed on NSL-KDD dataset.To reduce some elements of the attributes of this dataset in order to enhance the accuracy of analysis.The performance of Random Tree model,C4.5,ADTree RepTree,and Random Forest approaches of intrusion detection system will be applied.The increase in intrusion events over internet many network intrusion detection systems have been developed to prevent network attacks.