# mySolve: Matrix Inversion of the Hessian of the Log-Likelihood In glarma: Generalized Linear Autoregressive Moving Average Models

## Description

Inverts the second derivative matrix of the log-likelihood to obtain the estimated covariance matrix of the parameters.

## Usage

 `1` ```mySolve(A) ```

## Arguments

 `A` Matrix; the negative second derivative of the log-likelihood

## Details

`mySolve` attempts to invert its matrix argument. If the matrix supplied is not invertible, `ErrCode` is set to 1.

## Value

 `Ainv` inverse of the negative second derivative of the loglikelihood. If the inverse is unable to be obtained, returns the original negative second derivative of the log-likelihood. `ErrCode` Numeric; 0 if the inverse can be found, 1 if not.

## Author(s)

"William T.M. Dunsmuir" <w.dunsmuir@unsw.edu.au>

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21``` ```### Using the polio data data(Polio) y <- Polio[, 2] X <- as.matrix(Polio[, 3:8]) ## Construct the vectors of phi lags and theta lags theta.lags <- c(1, 2, 5) phi.lags <- rep(0, 0) ## Construct the initial delta vector delta <- c("Intcpt" = 0.2069383, "Trend" = -4.7986615 , "CosAnnual" = -0.1487333, "SinAnnual" = -0.5318768, "CosSemiAnnual" = 0.1690998, "SinSemiAnnual" = -0.4321435, "theta_1" = 0, "theta_2"= 0, "theta_5"= 0 ) ## Calculate the second derivative of the loglikelihood glarmamod <- glarmaPoissonPearson(y, X, delta = delta, phiLags = phi.lags, thetaLags = theta.lags, method = "FS") ## estimate the covariance matrix of the estimators from the second ## derivative of the loglikelihood mySolve(-glarmamod\$ll.dd) ```

glarma documentation built on May 2, 2019, 6:33 a.m.