Feature Selection Technique Using Principal Component Analysis for Improving Fuzzy C-Mean Internet Traffic Classification

Abstract

Background: Feature selection for within datasets has an important role in the process of internet traffic classification, feature selection in the presence of more precise data would make the internet traffic classification more accurate and can capable of providing more precise information for optimization of internet bandwidth. One of the important things in the feature selection technique is how to choose the discriminant feature which could in turn deliver better results during the classification process. Objective: Choose the discriminant features in internet traffic dataset using Principal Component Analysis ( PCA ) technique to improve classification accuracy Results: PCA technique improving the accuracy of Internet traffic classification using Fuzzy C – Mean, in this research the accuracyobtain 88.49 %and better than another feature selection technique Conclusion: PCAtechnique can be one of the solution for pre processing feature selection before internet traffic classification, because PCA can choose discriminant and principal feature in internet traffic classification dataset