1. Home
  2. Linear Mixed Models – Initial Values

Linear Mixed Models – Initial Values

For larger problems, where the data set is large or there are many model parameters to be estimated, the REML algorithm will run more efficiently if it is given good initial estimates of the variance parameters. This dialog allows you to specify a list of initial gamma ratios, which are the ratio of each component to the residual variance.

Current initial values

Lists the current initial values and constraints for each of the random model terms in the current model. Select the term that you want to change the details for.

Model term

Displays the name of the model term for which the current details are being displayed. Select a model term by choosing the appropriate name from the Current initial values list.

Initial value

Specify the initial gamma ratio. The default is 1. If you change this value then click Apply to activate the change.

Constraints

Lists the constraints that can be applied to the variance components. If Positive is selected then the variance component will be held positive. Selecting Fix will fix the component at its initial value. The option, None, will allow the variance component to become negative, provided the overall estimated variance-covariance matrix for the data remains positive definite.

See also

Linear Mixed Models (REML) 

Updated on April 18, 2019

Was this article helpful?