Abstract: To improve network security different steps has been taken as size and importance of the network has increases day by day. At that point opportunity of a system attacks increases. Network is for the most part assaulted by intrusions that are distinguished by system intrusion recognition framework. This paper attempts to build up an intrusion location framework which uses the signature and identity of the intrusion for distinguishing various types of intrusions. Here random forest tree algorithm was used for finding the patterns in the input data. In this work use of Gini index was done for the decision tree construction in recursive manner. Experiment was done on NSL-KDD dataset which was real. Comparison was done with latest RNN (Recurrent Neural Network). Result obtained after analyzing this system improved precision value by 12.06%, while recall value by 1.15% and accuracy values were improved by 6.87%.
Keywords: Clustering, Gini-Index, Intrusion Detection, Random Forest, Pattern generation
| DOI: 10.17148/IJARCCE.2019.8309