Modelling and Forecasting Nigeria’s Gross Domestic Product (GDP) with Hybrid ARIMA-RBF Model

Abstract

A nation's GDP is a crucial indicator of economic development and income levels. This study examines a hybrid model combining Autoregressive Integrated Moving Average (ARIMA) and Radial Basis Function (RBF) to best fit Nigeria's annual Gross Domestic Product (GDP) data from 1960 to 2022. The series was tested for stationarity using the Augmented Dickey-Fuller (ADF) test and was found to be stationary at the second differencing. The ARIMA (3, 2, 1) model was identified as the suitable ARIMA configuration.  Thereafter, RBF was used to model the nonlinearity in the residuals of the selected ARIMA. The best RBF model was the one consisting of one input with one neuron, one hidden layer with ten neurons and one output layer with one neuron N^(1,10,1). The results of the out-of-sample forecasted period showed that hybrid ARIMA-RBF outperformed the standalone ARIMA model.