MODELLING AND FORECASTING OF MILLET PRODUCTION IN NIGERIA
Keywords:
Millet Production, ARIMA, Model Evaluation Criteria, Forecasting, NigeriaAbstract
Millet is one of the most extensively cultivated cereals in the world, ranking fourth after rice, wheat
and sorghum. In Nigeria since independence it has been noticed that the area under millet
production has been gradually decreased, even though there is increase of productivity in some
areas. However, millet production hampered with numerous problems which drastically decreased
yields of millet production from about 8 million tonnes in the year 2008 to about 5 million, 1
million and even 0.9 million tonnes in 2009/2010, 2011/2012 and 2013 respectively. There is slight
increase in production of 2014, 2015, 2016 and 2017 of 1.3, 1.4, 1.5 and 1.5 million tonnes
respectively. There is need to examine the trend of millet production in Nigeria. Therefore, this
research modelled millet production in Nigeria for the period of 55 years from 1962 to 2017 using
the secondary data from Food and Agriculture of United Nations (FAO). The result shows that ARIMA
(1,1,0) is the model that best fit and describe the reality among the 10 postulated ARIMA models
because it has the least Akaike information criteria (AIC) and Bayesian information criteria (BIC),
the model evaluated its performance and forecast five years observations for 2018 to 2022, and the
results shows expected (increase in) millet production of 2,079,942; 2,522,333; 2,842,501;
3,074,214; and 3,241,909 tonnes for the year 2018; 2019; 2020; 2021; and 2022 respectively.
Finally, The ARIMA model is suggested for modelling millet production in Nigeria.