glmpred(variates,factors [, estimate:F, seest:F, sepred:T, n:N, silent:T]), variates and factors REAL vectors or matrices or NULL, N a positive scalar or REAL vector with positive elements. |

glmpred(Variates, Factors) computes estimates of the expected value of the response variable y for specified values of any variates and levels of any factors in the latest GLM model. It returns a structure with REAL components (vectors, except after manova()) "estimate" and "SEest". If there are no variates in the model, Variates should be NULL (see topic 'NULL'). Otherwise, Variates should be REAL. If there are Nvar variates in the model, Variates should either be a vector of length Nvar containing values for each of the variates, or a matrix with Nvar columns, with each row containing values for each variate. If there are no factors in the model, Factors should be omitted or explicitly NULL. Otherwise, Factors should be REAL. If there are Nfac variates in the model, Factors should either be a vector of length Nfac containing levels for each of the factors, or a matrix with Nfac columns, with each row containing levels for each factor. If either Variates or Factors contains data for only one case, it is used for all cases. Otherwise, you must have nrows(Variates) = nrows(Factors). Caution: After anova(), manova() and regress(), standard errors are computed using the final error mean square in the model. This may not be appropriate with mixed models, including split plot designs. glmpred(Variates, Factors, silent:T) does the same except that certain advisory messages are suppressed. 'silent:T' can be used with any other keywords. The default value of 'silent' is False unless the value of option' 'warnings' is False. glmpred(Variates, Factors, sepred:T) adds component SEpred to the output structure containing a vector or matrix of prediction standard errors. This is only permissible after regress(), anova() or manova() and their weighted alternatives. glmpred(Variates, Factors, seest:F) suppresses the computation of standard errors. glmpred(Variates, Factors, estimate:F) suppresses the computation of expected values. This option is legal only after anova(), manova(), regress() and their weighted alternatives. You cannot use glmpred() after fastanova() or ipf() or when coefs:F was used on the preceding GLM command. After GLM functions involving a Binomial response variable (logistic(), probit(), glmfit(...,dist:"binomial")), the values computed are the estimated probabilities p of "success" associated with each case (set of values). In this case, you can also use keyword phrase n:N, where N is a REAL variable, to specify the number of trials for each case. N can be a scalar or a vector whose length matches the number of cases. The resulting estimated values are N*phat, where phat are the estimated probabilities. After GLM functions such as poisson(), logistic(), or probit(), where the expectation of the response is a non-linear function of a linear combination of the predictors, the standard error is computed from the expectation and standard error in the linear scale using the delta-method. When the response is binomial and you also use n:N, the standard errors are those of N*p. Comment: Standard errors are computed on the assumption that all effects are fixed and not random. When this is not appropriate, the standard errors will usually indicate more precision than is warrented. Examples: After regress(), glmpred(x,sepred:T) is equivalent to regpred(). After anova("y=x+a+b"), x a variate, a and b factors, glmpred(x, hconcat(a,b)) computes fitted values and their standard errors for each case. glmpred(modelvars(variates:T), modelvars(factors:T)) computes fitted values and their standard errors for each case, regardless of the model. See also glmtable(), glmpred(), regpred(), modelinfo(), popmodel(), pushmodel().

Gary Oehlert 2003-01-15