Estimation of Types of States in Partial Observable Network Systems

Abstract

Estimating of types of states in network systems is essential for protecting network infrastructures from cyberattacks, managing network traffic, and detecting changes in network systems. It is very difficult to estimate the types of states in network systems due to their high complexity. The accuracy of the estimating the states in network systems depends heavily on the completeness of the collected sensor information. But the state of a network system at a given point in time may be never fully known due to noisy sensors; making more difficult to estimate the entire true state of a network system because certain features of the input data may be missing. In order to estimate the states in a network system in partially observable environments, an approach to estimating the types of states in partially observable cyber systems is presented. This approach involves the use of a convolutional neural network (CNN), and unsupervised learning with elbow method and k-means clustering algorithm.