Classification of unknown Internet traffic applications using Multiple Neural Network algorithm

Traffic classification software is an important tool in complex environments like a cloud-based environment for network and device safety. The new methods of traffic classification attempt to benefit from numerical flow characteristics and computer teaching techniques, but minimal supervised knowledge and uncertain applications seriously affect classification efficiency. We propose a new way of dealing with an unknown application issue in the critical situation of a limited supervised training set to achieve an efficient network classification. The proposed model applied the multiple neural network algorithms to predict the unknown application that run through organization internet network. The advantage of the suggested approach is to filter and exclude the unknown internet applications that can be affecting into internet network performance. By Appling proposed method, the internet performance can be improved and the internet traffic and delay of transferred data can be reduced. The proposed method compared with other based line method in term of predication precision accuracy measure.