predict.lmm: Predicted Mean Value With Uncertainty For Linear Mixed Model In LMMstar: Repeated Measurement Models for Discrete Times

Description

Predicted mean value conditional on covariates or on covariates and other outcome values.

Usage

 ``` 1 2 3 4 5 6 7 8 9 10 11 12``` ```## S3 method for class 'lmm' predict( object, newdata, se = "estimation", df = !is.null(object\$df), type = "static", level = 0.95, keep.newdata = FALSE, se.fit, ... ) ```

Arguments

 `object` a `lmm` object. `newdata` [data.frame] the covariate values for each cluster. `se` [character] Type of uncertainty to be accounted for: estimation of the regression parameters (`"estimation"`), residual variance (`"residual"`), or both (`"total"`). Can also be `NULL` to not compute standard error, p-values, and confidence intervals. `df` [logical] Should a Student's t-distribution be used to model the distribution of the predicted mean. Otherwise a normal distribution is used. `type` [character] Should prediction be made conditional on the covariates only (`"static"`) or also on outcome values at other timepoints (`"dynamic"`). Can also output the model term (`"terms"`, similarly to `stats::predict.lm`. `level` [numeric,0-1] the confidence level of the confidence intervals. `keep.newdata` [logical] Should the argument `newdata` be output along side the predicted values? `se.fit` For internal use. When not missing mimic the output of `predict.se`. Overwrite argument `se`. `...` Not used. For compatibility with the generic method.

Details

Static prediction are made using the linear predictor while dynamic prediction uses the conditional normal distribution of the missing outcome given the observed outcomes. So if outcome 1 is observed but not 2, prediction for outcome 2 is obtain by X_2β + σ_{21}σ^{-1}_{22}(Y_1-X_1β). In that case, the uncertainty is computed as the sum of the conditional variance σ_{22}-σ_{21}σ^{-1}_{22}σ_{12} plus the uncertainty about the estimated conditional mean (obtained via delta method using numerical derivatives).

The model terms are computing by centering the design matrix around the mean value of the covariates used to fit the model. Then the centered design matrix is multiplied by the mean coefficients and columns assigned to the same variable (e.g. three level factor variable) are summed together.

Value

A data.frame with 5 columns:

• `estimate`: predicted mean.

• `se`: uncertainty about the predicted mean.

• `df`: degree of freedom

• `lower`: lower bound of the confidence interval of the predicted mean

• `upper`: upper bound of the confidence interval of the predicted mean

except when the argument `se.fit` is specified (see `predict.lm` for the output format).

Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22``` ```## simulate data in the long format set.seed(10) dL <- sampleRem(100, n.times = 3, format = "long") ## fit Linear Mixed Model eUN.lmm <- lmm(Y ~ visit + X1 + X2 + X5, repetition = ~visit|id, structure = "UN", data = dL) ## prediction newd <- data.frame(X1 = 1, X2 = 2, X5 = 3, visit = factor(1:3, levels = 1:3)) predict(eUN.lmm, newdata = newd) predict(eUN.lmm, newdata = newd, keep.newdata = TRUE) ## dynamic prediction newd.d1 <- cbind(newd, Y = c(NA,NA,NA)) predict(eUN.lmm, newdata = newd.d1, keep.newdata = TRUE, type = "dynamic") newd.d2 <- cbind(newd, Y = c(6.61,NA,NA)) predict(eUN.lmm, newdata = newd.d2, keep.newdata = TRUE, type = "dynamic") newd.d3 <- cbind(newd, Y = c(1,NA,NA)) predict(eUN.lmm, newdata = newd.d3, keep.newdata = TRUE, type = "dynamic") newd.d4 <- cbind(newd, Y = c(1,1,NA)) predict(eUN.lmm, newdata = newd.d4, keep.newdata = TRUE, type = "dynamic") ```

LMMstar documentation built on Nov. 5, 2021, 1:07 a.m.