Table 4 shows the least-squares characteristics related to the regression of sorghum production onto population and precipitation.
Table 4. Least squares statistics associated with the regressions of Sorghum
3.3.3. Maize
From the analysis using MATLAB, the best model fit to maize data was the sum of sine model with three terms and the adjusted R2 of 82.96% and the RMSE of 2.882x104. We can see that maize production has been increasing while oscillating with three local peaks around years 1981, 1994 and 2008 (Figure 6.a). The variations in maize production can be explained by the variations in population (Figure 6.b). The sum of sine function is translated to f=( a1*sin(b1*(x+3)+c1) + a2*sin(b2*(x+3)+c2) + a3*sin(b3*(x+3)+c3)), which means that there was a delay of 2 to 3 years (see supplemental IV for details). In particular, the changes in maize production occurred two-three years after the variations below or above the trend of population increase. Hence, the variations in maize production is a function of the variations in the de-trend population data with a delay of 2 to 3 years.