Prediction of macroeconomic variables of Pakistan: Combining classic and artificial network smoothing methods

Abstract

This research concentrates on using neural networks in the modelling and prediction of macroeconomic variables in specific. Macroeconomic predictors are particularly interested in neural networks because of their capacity to predict any linear or non-linear connection with a decent degree of precision. Two macroeconomic variables have been used for projecting: gross domestic product (volume, NGDPD) and total investment (NID NGDP) over the time period of 2013–2023. Moving averages, exponential smoothing, Brown’s single-parameter linear, exponential smoothing, Brown’s second-order exponential smoothing, Holt’s two-parameter linear exponential smoothing, and decomposition techniques are used as analytical tools. The research focuses on the usefulness of the artificial neural network model for predicting economic determinants in the long run and compares the ANN’s findings with the Conventional Time - Series data sets (Smoothing & Decomposition Techniques). To emphasise the point, a scientific illustration is used to forecast Pakistan’s two crucial macroeconomic indicators. Based on the empirical results, ANN can play a vital role in forecasting the macroeconomic fundamentals of Pakistan if compare to Exponential smoothing techniques.