
207Formulating the Required Equations
β
0
is taken to be the value that maximizes the log-likelihood. In this simple
case, the maximum can be easily seen by graphing the likelihood function,
as illustrated in Figure6.6. In typical cases where the model involves mul-
tiple parameters, the set of values that together maximize the likelihood is
taken as the best estimate. These values are found mathematically (either
in closed form solutions) or using numerical procedures that search for the
maximum. The likelihood approach is broadly used due to the properties of
the estimates obtained this way; in particular, with large enough samples,
the para ...