A NOVEL APPROACH FOR IMAGE IMPAINTING USING ADAPTIVE SVM

Abstract

Data security is critical for most businesses and even home computer users. As, all the sensitive and important information is manipulated online. Therefore, it is essential to protect the data from intruders. We need to find the best ways to protect our system from different types of attacks.Intrusions refer to the network attacks against vulnerable services, data-driven attacks on applications, host-based attacks like privilege escalation, unauthorized logins and access to sensitive files, or malware like viruses, worms and Trojan horses. The problem in current IDS is the high false positive rate and low detection rate.As far as system complexity is concern clustering algorithm having higher space and time complexity.Moreover taking consider all the above mentioned limitation there is requirement of novel data mining technique which can give better accuracy, system should make automated self-configured self-optimized and having least complexity and very quick response time as we are taking into consideration the IDS system which is very important aspect of cloud computing, also it should not make system heavier, does not affect the QOS, and so on, hierarchical clustering seems to be better solution for all the problem also in future we try to add some classification technique in this system so that would make fully automated system.