Autoregressive Model for Cocoa Production in Nigeria

Abstract

In this study, the trend and stationarity of cocoa production was examined to check whether it satisfied the statistical assumptions before the autoregressive model of order two after first difference was selected. The plot of the cocoa production was not stationary in mean, as the level changes over time. Phillips-Perron Unit Root Test was used to check the claim and equally showed that cocoa production was not stationary. There was very little evidence for non-zero autocorrelations in the forecasts errors at lags 1-20.The forecast errors seems to have roughly constant variance over time. The histogram of the time series showed that the forecast errors were roughly normally distributed and the mean seems to be close to zero. Therefore, it was plausible that the forecast errors were normally distributed with mean zero and constant variance. The R statistical package was used for all the analyses in this study.