Prediction of chicken prices during Covid-19 pandemic using VAR, Kernel, and Fourier series simultaneously

Abstract

One of the goals of the Sustainable Development Goals (SDGs) is to achieve good food security. However, this goal is difficult to implement due to the Coronavirus Disease 2019 (Covid-19). One of the impacts of the Covid-19 pandemic on the trade sector is the change in prices of several main commodities, such as chicken meat and eggs. Firstly, this study uses the Vector Autoregressive (VAR) to predict the prices of chicken meat and eggs. However, there are several parameters that are not significant and the assumptions of data stationarity, residual simultaneous normality, and residual homogenity are not met. Thus, simultaneous nonparametric methods, that is the kernel and Fourier series, is used to predict the prices of chicken commodity. Simultaneous kernel modeling produces a Gaussian function with h = 0.65 as the best kernel function, while simultaneous Fourier series produces a cosine sine function with γ and π. The Fourier series produces K= 119 as the best function. So, simultaneous Gaussian-kernel model is the best model based on the criteria of Root Mean Square Error (RMSE) and R2, with the value of 107.93 and 99.83% for chicken meat, and 16.54 and 99.97% for chicken eggs, respectively. The best model has good performance in prediction with the Mean Absolute Percentage Error (MAPE) value for chicken meat price of 3.2444%, while for chicken egg price of 3.758%. The prediction results of the simultaneous Gaussian-kernel model are expected to be a reference for the government in controlling related commodity prices during the Covid-19 pandemic. © 2022 American Institute of Physics Inc.. All rights reserved.