Improved Intrusion Detection in DDoS Applying Feature Selection Using Rank & Score of Attributes in KDD-99 Data Set

In today´s networked environment, massive volume of data being generated, gathered and stored in databases across the world. This trend is growing very fast, year after year. Today it is normal to find databases with terabytes of data, in which vital information and knowledge is hidden. The unseen information in such databases is not feasible to mine without efficient mining techniques for extracting information. In past years many algorithms are created to extract knowledge from large sets of data. There are many different methodologies to approach data mining: classification, clustering, association rule, etc. Classification is the most conventional technique to analyse the large data sets. Classification can help identify intrusions, as well as for discovering new and unknown types of intrusions. For classification, feature selection provides an efficient mechanism to analyse the dataset. We are trying to analyse the NSL-KDD cup 99, dataset using various classification algorithms. Primary experiments are performed in WEKA environment. The accuracy of the various algorithms is also calculated. A feature selection method has been implemented to provide improved accuracy. The main objective of this analysis is to deliver the broad analysis feature selection methods for NSL-KDD intrusion detection dataset.