Network Vector Autoregressive Moving Average Model

In this paper, we propose a NARMA model that integrates both the autoregressive and moving average components into the network models. We provide definitions for strict stationarity and invertibility, along with a closed-form solution, and identify the corresponding conditions for the analytical solution to be stationary and reversible. Through numerical simulations and empirical experiments on the S&P 500 index, we demonstrate that our proposed model is flexible, accurate, and easy to use in modeling network structures, while also addressing the deficiencies in previous literature related to modeling the moving average component.