Table 1. Model goodness of fit comparisons using a selection of stochastic differential equation and linear regression models.
| Model | No.
parameters |
Remarks | AIC
(smaller is better) |
BIC
(smaller is better) |
|---|---|---|---|---|
| Ornstein-Uhlenbeck
( Equation 3) |
3 | No random effects | 12181 | 12201 |
| Linear regression | 15454 | 15473 | ||
| Ornstein-Uhlenbeck
with mixed effects |
6 | Random speed of reversion
1
and long-term mean with covariance 2 |
11544 | 11564 |
| Linear mixed effects
regression |
Random intercept and slope
with covariance 2 |
11607 | 11627 |
AIC, Akaike information criterion; BIC, Bayesian information criterion. . 1α in the Wiener (Brownian motion) process and is the covariance between time points (see text above for detail); 2Child-level random effects.