This class of objects is returned from the nlme function to represent a fitted nonlinear mixed effects model. Objects of this class have methods for the generic functions anova, coef, fitted, fixed.effects, plot, predict, print, random.effects, residuals, and summary. Objects from this class inherit from class lme.

---

VALUE:

The following components must be included in a legitimate nlme object. The residuals, fitted values and coefficients can be extracted by the generic functions of the same name, or by the "$" operator.

coefficients:

a list with two components, fixed and random, where the first is a vector containing the estimated coefficients for the fixed effects and the second is a matrix containing the estimated coefficients for the random effects. The columns refer to the parameters in the random formula, and the rows to the cluster levels; the names of the coefficients are the same as those in the fixed or random formulas when just a . is used on the right hand side, or the original name with the covariate name appended to it (i.e. name.covariate) otherwise.

fitted.values:

a data frame containing the population and cluster fitted values. The population fitted values are evaluated at the converged estimates of the fixed effects and the mean of the random effects (That is the random effects are set to zero). The cluster fitted values are evaluated at the converged estimates of the fixed effects and the conditional estimates of the random effects.

residuals:

a data frame containing the population and cluster residuals from the fit. The population residuals are the observed values minus the population fitted values and the cluster residuals are the observed values minus the cluster fitted values. If the variance function structure used in the fit is different than "identity", a column with the estimated individual standard deviations will be included in the residuals data frame.

var.ran:

Random effects variance-covariance-correlation matrix estimate. The variances estimates for the random effects are displayed on the main diagonal, covariances above the diagonal and correlations below the diagonal.

var.fix:

the estimated variance-covariance matrix of the fixed effects.

sigma:

estimate of the cluster residual standard deviation.

loglik:

the approximate log likelihood at convergence.

cluster:

a vector containing the clusters labels.

call:

a list containing an image of the nlme call.

nobs:

nclus:

the number of observations (nobs) and clusters (nclus) used in the fit.

est.method:

estimation method used in the fit, either "ML" or "RML".

niter:

number of iterations used in iterative algorithm.

re.block:

re.structure:

the blocking and covariance structures used for the random effects.

re.paramtr:

the parametrization used for the unstructured variance-covariance matrices of the blocks of random effects.

sizetheta:

length of the parameter vector theta used to obtain the scaled variance-covariance matrix of the random effects D.

criterion:

a vector containing the values of the convergence criteria for the fixed effects, the factor of the scaled covariance matrix of the random effects, and optionally the serial structure and variance function parameters.

serial.structure:

the variance-covariance structure assumed for the within-cluster errors.

alpha:

the estimated serial structure parameter vector, if serial.covariate is not equal to "identity".

var.function:

the variance function structure used for the within-cluster errors.

delta:

the estimated variance function parameters, if var.function is not equal to "identity".

---

SEE ALSO:

nlme , nlme.formula , nlme.nlsList .