NETWORK SECURITY USING PATTERN BASED NEURAL NETWORK

Abstract

The present research relates to the computer system user authentication, identification and security in a small to large multi-user environment. This presented program supports the automatic system administration to limit the access of the unauthorized person/program by tracking the uses of application software pattern in a network. In this program if suspicious program/activity/commands are happen then this limit the unauthorized person/program to disrupt system operations by using a neural network and set of rules to track the application software uses pattern and mark/identify suspicious activity on the system.

Through this research we not only remove the dependency of user name and password for the user authentication but also track the suspicious activity within the network and prevent the user/network/company/organization through hacking .once we start the program in a network the neural network automatically train itself for the rapidly used application software patterns and We train the neural network beyond the tasks. Once we limit the tasks user never cross the boundaries it all depends on neural network’s learning process and duration of learning period where we teach the neural network in future we fix the work deadlines

The main objective of this research work is to provide support to system administrators in limiting the ability of unauthorized user/ abnormal activities by monitoring the application software usage pattern and decision taking by the neural network.

The present research work is a method that will prevent destructive programs, malicious software/ hacking tools from being executed. A combination of security rules and application software usages pattern are used to identify suspicious activities on a network/computer.