# muSigma: Function to estimate mean and covariance for unknown... In indirect: Elicitation of Independent Conditional Means Priors for Generalised Linear Models

## Description

Function to estimate mean and covariance for unknown parameters β.

## Usage

 `1` ```muSigma(Z, X = NULL, fit.method = "KL", wls.method = "default") ```

## Arguments

 `Z` list of design points and link function that is an output of function `designLink` `X` model matrix for model formula and design points. The covariates must correspond to the description of design points in `Z`, but can be transformed etc. If `NULL` then `X` will be coerced by applying `as.matrix()` to `Z\$design`. The matrix `X` should be full rank when subsetted to the elicited design points. If a column of `X` has the name `offset` then this column is treated as an offset during estimation `fit.method` character, `moment`, `KL`. See `mV`. Default is `KL`. `wls.method` character giving the numerical solution method: `QR`, using the QR decomposition, `SVD`, using the singular value decomposition, or option `default` that uses `solve()`

## Value

list of `mu`, numeric vector of location parameters for the normal prior; `Sigma`, the covariance matrix; and `log.like`, a scalar

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15``` ```X <- matrix(c(1, 1, 0, 1), nrow = 2) # design Z <- designLink(design = X) Z <- elicitPt(Z, design.pt = 1, lower.CI.bound = -1, median = 0, upper.CI.bound = 1, comment = "The first completed elicitation scenario.") Z <- elicitPt(Z, design.pt = 2, lower.CI.bound = -2, median = 1, upper.CI.bound = 2, comment = "The second completed elicitation scenario.") prior <- muSigma(Z, X, fit.method = "KL") prior\$mu prior\$Sigma ```

indirect documentation built on May 1, 2019, 6:35 p.m.