Comparative study of multivariate and univariate models in forecasting time series data

Abstract

This work compares the performances of univariate and multivariate time series models. Five time series variables from Nigeria’s gross domestic products were used for the comparative study. These series were modelled using both the univariate and multivariate time series framework. The performances of the two methods were evaluated based on the mean error incurred by each approach. The results showed that the univariate linear stationary models perform better than the multivariate models.