Table 2.

The four predictive models

Model	Equation	df
Linear	Nx = a + bx + ε	1
Quadratic	Nx = c + dx + ex2 + ε	2
Cube polynomial	Nx = f + gx + hx2 + jx3 + ε	3
Exponential	Nx = kexp(mx) + ε	1

where a, c, f and k are intercepts; b, d, e, g, h, j and m are slopes which are estimated by least square estimation, x is the time-point and ɛ’s are mutually uncorrelated random errors with mean (0) and common variance (σ²); df is the degree of freedom