WSEAS Transactions on Computers
Print ISSN: 1109-2750, E-ISSN: 2224-2880
Volume 14, 2015
An Improved Neural Network Based Approach for Identification of Self & Non-Self Processes
Authors: ,
Abstract: Security of computer systems is a very crucial issue. Now a days various security approaches and tools are used to protect computer system by any virus, worms and attacks. These security tools require a regular signature based updating to protect the computer system by latest virus and worms. If a system has not updated its security tool, then it may be infected by any virus or worms, then the operating system generates its processes. These processes are harmful to the computer system. These processes are categorized as non-self processes. In this paper an Artificial Neural Network is designed to identify the self and non-self operating system process. Backpropagation Algorithm is used to provide the training and learning to the Artificial Neural Network. Initially an Artificial Neural Network is created with random input weights. These weights are updated by using Backpropagation Algorithm for various training examples. After the weight update Artificial Neural Network tests by various test data examples. After campaigning with various computer security approaches it has been observed out, that Artificial Neural Network provides a better security by identifying self and non-self process.
Search Articles
Pages: 272-286
WSEAS Transactions on Computers, ISSN / E-ISSN: 1109-2750 / 2224-2880, Volume 14, 2015, Art. #28