FUZZY MODEL FOR TIME SERIES FORECASTING

Abstract: In 2007, in Kazakhstan, there was a transition of TDM (Time Division Multiplexing) circuit-switched technologies to IP (Internet Protocol) packet technology, which created modern infrastructure for the ICT (information communication technologies) sphere. The advent IoT Things) concept has led growth functioning network at faster rate. It is currently developing direction cognitive infocommunication network. Its evolutionary development characterized by change volume transmitted information, types its presentation, methods transmission and storage, number sources consumers, distribution among users, requirements timeliness reliability (quality) [1]. Types traffic their structure are changing, therefore data processing becomes more complicated. For this reason, tasks analyzing predicting remain relevant.
 work, prediction measured on real performed. series under study shows totality packets over backbone each second. Forecasting one-dimensional time carried out basis fuzzy logic methods. This class models well suited modeling nonlinear systems forecasting. use sets based ability approximate functions, as readability rules using linguistic variables. results software algorithm inference were obtained Python environment. Membership predictive graphs built evaluation out. numerical values root mean square error (MSE) calculated. As result, it found that Cheng model higher forecast accuracy than Chen forecasting method.