Identification of attacks using proficient data interested decision tree algorithm in data mining

Abstract

The key feature of today’s networks is being open, communication between any pair of Internet end points is easier. This leads to various types of intrusions which are actions that threaten the confidentially of the network and lack of effective network infrastructures for distinguishing and dropping malicious traffics. This approach of intrusion detection with data mining concepts involving the KDD cup dataset that generates rules for the detection which works well for new as well as unknown attacks. Data mining is the process of identifying valid understandable patterns in data. It can help learn the traffic through supervised and unsupervised learning we have applied here the semi supervised way. To classify the given data resourcefully, the Proficient Data Interested Decision Tree (PDIDT) algorithm is functioned. We have concentrated on mitigating the Distributed Denial of service (DDos) attacks and in reducing the false alarm rate (FAR) with a global network monitor which can observe and control every flow between any pair of hosts.