OUTLIER DETECTION AND SYSTEM ANALYSIS USING MINING TECHNIQUE OVER KDD

ABSTRACT The intrusion detection system has been implemented using various data mining techniques which help user to identify or classify various attacks or number of intrusion in a network. KDD dataset is one of the popular dataset to test classification technique s. In this paper our work is done on analysis of different techniques which were used in order to detect outlier and get the IDS system improved based on the system. The paper investigate the algorithm such as ? Naïve Baise, FPOP, SCF, K-Mean and A hybrid approach of Genetic and SVM which is combined and proposed by us to find better outlier as compare to other proposed algorithm by different authors. Our work contribute the investigation and analysis of different outlier detection technique over KDD dataset.