New study finds better approximations for autoregressive models with non-zero initial conditions.
The article explores different ways to estimate a mathematical model when the starting point is not zero. The researchers compared three methods and found that one method, Perron's continuous time approximation, works best when the initial condition is not zero.
