make.design {surveillance} | R Documentation |
Create the design matrices
Description
Creates the design matrices needed for meanResponse
Usage
make.design(disProgObj, control=list(linear=FALSE, nseason=0, period=52,
neighbours=FALSE, negbin=FALSE, lambda=TRUE))
Arguments
disProgObj |
Object of class disProg |
control |
Control object:
- linear
- if
true a linear trend is included.
- nseason
- number of Fourier frequencies to be used
- period
- period length (defaults to 52 for weekly data)
- neighbours
- if
true a autoregressive parameter for
adjacent areas is included.
- negbin
- if
true negative binomial rather than poisson is used.
- lambda
- if
true a autoregressive parameter is included.
|
Details
Value
list |
- Y
- matrix with number of cases y_{it} in unit i at
time t as elements, i.e. data without the first time point.
- Ym1
- matrix with previous number of cases y_{i,t-1},
i.e data without the last time point.
- Ym1.neighbours
- matrix with sum of earlier counts of adjacent units
sum_{j sim i} y_{j,t-1}
- X.trendSeason
- design matrix for linear trend and seasonal components
- X.predict
- design matrix for prediction
- populationFrac
- matrix with corresponding population proportions
- dimtheta
- number of parameters used in model
- psiIndex
- position of psi in vector with parameters
- control
- control object
|
Author(s)
M.Paul, L. Held
Source
See Also
Examples
[Package
surveillance version 0.9-6
Index]