Internet Traffic Prediction Model

Abstract

Many modern machine learning tools are inefficient due to the pronounced nonlinearity of traffic changes and nonstationarity. For this, the task of predicting the signs of increments (directions of change) of the process of time series is singled out. This article proposes the use of some results of the theory of random processes for a quick assessment of the predictability of signs of increments with acceptable accuracy. The proposed procedure is a simple heuristic rule for predicting the increment of two neighboring values for a random sequence. The connection of this approach to time series with known approaches to the prediction of binary sequences is shown. The possibility of using the experience of predicting the absolute values of traffic in predicting the signs of changes is considered.