Publication

Abstract : The domain of network traffic and anomaly reporting has its highlights in recent days towards real-time applications. The utilization of network data with its correlating attributes and its measures must be considered before deriving a pattern or decision making process. SDN(Software Define Networks) in turn has been facilitating more towards the efficient and reliable mechanism for data control and network management system. This research focus on the development of a hybrid prediction model for disseminating the factors that contribute towards anomaly detection in SDN. The experimentation can be made by hybridizing different classification schemes with tuned parametric components upon statistical modelling and evaluation. The proposed hybrid model is suggested up with the hybridization of SVM, NN along with Ada Boost algorithm with tuned parameter metrics. From the proposed model it is evident that the selected features will play a significant key role in anomaly prediction in a given network.