NAG CL Interface
g02jfc (lmm_​init)

Settings help

CL Name Style:


1 Purpose

g02jfc preprocesses a dataset prior to fitting a linear mixed effects regression model via g02jhc.

2 Specification

#include <nag.h>
void  g02jfc (void **hlmm, void *hddesc, void *hfixed, Integer nrndm, void * hrndm[], Integer n, const double y[], const double wt[], const double dat[], Integer pddat, Integer sddat, Integer *fnlsv, Integer *nff, Integer *rnlsv, Integer *nrf, Integer *nvpr, double rcomm[], Integer lrcomm, Integer icomm[], Integer licomm, NagError *fail)
The function may be called by the names: g02jfc or nag_correg_lmm_init.

3 Description

g02jfc must be called prior to fitting a linear mixed effects regression model via g02jhc.
The model is of the form:
y=Xβ+Zν+ε  
where y is a vector of n observations on the dependent variable,
X is an n×p design matrix of fixed independent variables,
β is a vector of p unknown fixed effects,
Z is an n×q design matrix of random independent variables,
ν is a vector of length q of unknown random effects,
ε is a vector of length n of unknown random errors.
Both ν and ε are assumed to have a Gaussian distribution with expectation zero and variance/covariance matrix defined by
Var[ ν ε ] = [ G 0 0 R ]  
where R=σR2I, I is the n×n identity matrix and G is a diagonal matrix. It is assumed that the random variables, Z, can be subdivided into gq groups with each group being identically distributed with expectation zero and variance σi2. The diagonal elements of matrix G, therefore, take one of the values {σi2:i=1,2,,g}, depending on which group the associated random variable belongs to.
The model, therefore, contains three sets of unknowns: the fixed effects β, the random effects ν and a vector of g+1 variance components γ, where γ={σ12,σ22,, σ g-1 2 ,σg2,σR2}.
Case weights can be incorporated into the model by replacing X and Z with Wc1/2X and Wc1/2Z respectively where Wc is a diagonal weight matrix.
The design matrices, X and Z, are constructed from an n×md data matrix, D, a description of the fixed independent variables, Mf, and a description of the random independent variables, Mr. See Section 11 for further details.

4 References

Rao C R (1972) Estimation of variance and covariance components in a linear model J. Am. Stat. Assoc. 67 112–115
Wolfinger R, Tobias R and Sall J (1994) Computing Gaussian likelihoods and their derivatives for general linear mixed models SIAM Sci. Statist. Comput. 15 1294–1310

5 Arguments

1: hlmm void ** Input/Output
On entry: must be set to NULL or, alternatively, an existing G22 handle may be supplied in which case g02jfc will destroy the supplied G22 handle as if g22zac had been called.
On exit: holds a G22 handle to the internal data structure containing a description of the model. You must not change the G22 handle other than through the functions in Chapters G02 or G22.
2: hddesc void * Input
On entry: a G22 handle to the internal data structure containing a description of the data matrix, D, as returned in hddesc by g22ybc.
3: hfixed void * Input
On entry: a G22 handle to the internal data structure containing a description of the fixed part of the model Mf as returned in hform by g22yac.
If hfixed is NULL then the model is assumed to not have a fixed part.
4: nrndm Integer Input
On entry: the number of elements used to describe the random part of the model.
Constraint: nrndm0.
5: hrndm[nrndm] void * Input
On entry: a series of G22 handles to internal data structures containing a description of the random part of the model Mr as returned in hform by g22yac. If nrndm=0, hrndm is not referenced and may be NULL.
6: n Integer Input
On entry: n, the number of observations in the dataset, D.
Constraint: 1nnd, where nd is the value supplied in nobs when hddesc was created.
7: y[n] const double Input
On entry: y, the vector of observations on the dependent variable.
Constraint: y[i-1]0.0 for at least one i=1,2,,n.
8: wt[n] const double Input
On entry: optionally, the diagonal elements of the weight matrix Wc.
If wt[i-1]=0.0, the ith observation is not included in the model and the effective number of observations is the number of observations with nonzero weights.
If weights are not provided then wt must be set to NULL, and the effective number of observations is n.
Constraint: if wtis notNULL, wt[i-1]0.0, for i=1,2,,n
9: dat[pddat×sddat] const double Input
Note: the (i,j)th element of the matrix is stored in dat[(j-1)×pddat+i-1].
On entry: the data matrix, D. By default, Dij, the ith value for the jth variable, for i=1,2,,n and j=1,2,,md, should be supplied in dat[(j-1)×pddat+i-1].
If the optional parameter Storage Order, described in g22ybc, is set to VAROBS, Dij should be supplied in dat[(i-1)×pddat+j-1].
If either yi, wi or Dij, for a variable j used in the model, is NaN (Not A Number) then that value is treated as missing and the whole observation is excluded from the analysis.
10: pddat Integer Input
On entry: the stride separating matrix row elements in the array dat.
Constraints:
  • if the optional parameter Storage Order, described in g22ybc, is set to VAROBS, pddatmd;
  • otherwise pddatn.
11: sddat Integer Input
On entry: the secondary dimension of dat.
Constraints:
  • if the optional parameter Storage Order, described in g22ybc, is set to VAROBS, sddatn;
  • otherwise sddatmd.
12: fnlsv Integer * Output
On exit: the number of levels for the overall subject variable in Mf. If there is no overall subject variable, fnlsv=1.
13: nff Integer * Output
On exit: the number of fixed effects estimated in each of the fnlsv subject blocks. The number of columns, p, in the design matrix X is given by p=nff×fnlsv.
14: rnlsv Integer * Output
On exit: the number of levels for the overall subject variable in Mr. If there is no overall subject variable, rnlsv=1.
15: nrf Integer * Output
On exit: the number of random effects estimated in each of the rnlsv subject blocks. The number of columns, q, in the design matrix Z is given by q=nrf×rnlsv.
16: nvpr Integer * Output
On exit: g, the number of variance components being estimated (excluding the overall variance, σR2). This is defined by the number of terms in the random part of the model, Mr (see Section 11 for details).
17: rcomm[lrcomm] double Communication Array
On exit: a communication array as required by the functions g02jgc or g02jhc.
18: lrcomm Integer Input
On entry: the dimension of the array rcomm.
19: icomm[licomm] Integer Communication Array
On exit: a communication array as required by the functions g02jgc or g02jhc.
If licomm or lrcomm are too small and licomm2, then fail.code= NE_ARRAY_SIZE and icomm[0] holds the minimum required value for licomm and icomm[1] holds the minimum required value for lrcomm.
20: licomm Integer Input
On entry: the dimension of the array icomm.
21: fail NagError * Input/Output
The NAG error argument (see Section 7 in the Introduction to the NAG Library CL Interface).

6 Error Indicators and Warnings

NE_ALLOC_FAIL
Dynamic memory allocation failed.
See Section 3.1.2 in the Introduction to the NAG Library CL Interface for further information.
NE_ARRAY_SIZE
On entry, licomm=value and lrcomm=value.
Constraint: licommvalue and lrcommvalue. icomm is not large enough to hold the minimum array sizes.
On entry, md=value and pddat=value.
Constraint: pddatmd.
On entry, md=value and sddat=value.
Constraint: sddatmd.
On entry, n=value and pddat=value.
Constraint: pddatn.
On entry, n=value and sddat=value.
Constraint: sddatn.
NE_BAD_PARAM
On entry, argument value had an illegal value.
NE_FIELD_UNKNOWN
A variable name used when creating hfixed is not present in hddesc.
Variable name: value.
A variable name used when creating hrndm is not present in hddesc.
Variable name: value.
NE_HANDLE
hddesc has not been initialized or is corrupt.
hddesc is not a G22 handle as generated by g22ybc.
hfixed has not been initialized or is corrupt.
hfixed is not a G22 handle as generated by g22yac.
i=value.
hrndm[i-1] has not been initialized or is corrupt.
i=value.
hrndm[i-1] is not a G22 handle as generated by g22yac.
On entry, hlmm is not NULL or a recognised G22 handle.
NE_INT
On entry, n=value.
Constraint: n1.
On entry, n=value and nd=value.
Constraint: nnd, where nd is the value supplied in nobs when hddesc was created.
On entry, no observations due to zero weights or missing values.
On entry, nrndm=value.
Constraint: nrndm0.
NE_INTERNAL_ERROR
An internal error has occurred in this function. Check the function call and any array sizes. If the call is correct then please contact NAG for assistance.
See Section 7.5 in the Introduction to the NAG Library CL Interface for further information.
NE_NO_LICENCE
Your licence key may have expired or may not have been installed correctly.
See Section 8 in the Introduction to the NAG Library CL Interface for further information.
NE_REAL_ARRAY
On entry, column j of the data matrix, D, is not consistent with information supplied in hddesc, j=value.
On entry, i=value and wt[i-1]=value.
Constraint: wt[i-1]0.0.
NE_ZERO_VARS
No model has been specified.
NW_ARRAY_SIZE
On entry, licomm=value and lrcomm=value.
Constraint: licommvalue and lrcommvalue. The minimum array sizes for licomm and lrcomm are held in the first two elements of icomm repectively.
NW_POTENTIAL_PROBLEM
Column j of the data matrix, D, required rounding more than expected when being treated as a categorical variable, j=value.
All output is returned using the rounded value(s).
The fixed part of the model contains categorical variables, but no intercept or main effects terms have been requested.

7 Accuracy

Not applicable.

8 Parallelism and Performance

g02jfc makes calls to BLAS and/or LAPACK routines, which may be threaded within the vendor library used by this implementation. Consult the documentation for the vendor library for further information.
Please consult the X06 Chapter Introduction for information on how to control and interrogate the OpenMP environment used within this function. Please also consult the Users' Note for your implementation for any additional implementation-specific information.

9 Further Comments

None.

10 Example

This example fits a random effects model with three random submodels and two fixed effects to a simulated dataset with 90 observations and 12 variables. The model is fit using maximum likelihood (ML). Standard labels for the parameter estimates and variance components are obtained from g22ydc. See g02jhc for an example of how to construct custom labels.

10.1 Program Text

Program Text (g02jfce.c)

10.2 Program Data

Program Data (g02jfce.d)

10.3 Program Results

Program Results (g02jfce.r)

11 Algorithmic Details

11.1 Fixed Effects Design Matrix, X

The fixed effects design matrix, X, is constructed from the data matrix D and Mf, as encoded in hfixed. Details of the construction are described in Section 3 in g22yac and Section 3 in g22ycc.
It is possible to store the cross-product matrix, XTX in a block diagonal form if Mf contains an overall subject effect, Sf. In this context Sf is defined as a main effect or interaction term that is contained in all other terms. For example, if Mf simplifies to V1.V4+V1.V2.V4+V1.V2.V3.V4, then Sf=V1.V4. If it is advantageous to do so, g02jfc will make use of this block diagonal structure and fnlsv will be set to the number of levels in Sf, otherwise fnlsv=1.

11.2 Random Effects Design Matrix, Z

The random effects design matrix, Z, is constructed from the data matrix D and Mr which is made up of nrndm submodels, Mri, where Mri is encoded in hrndm[i-1]. Each submodel is made up of two parts, the random effects and a subject term. The random effects are specified as described in Section 3 in g22yac and the subject term is specified via the g22yac optional parameter Subject. The design matrix Z is constructed as described in Section 3 in g22ycc using a model constructed from the nrndm submodels. As an example, if there were 3 submodels: then Z would be constructed as if g22ycc was called using the model It should be noted that unless specified otherwise (by the inclusion of -1) a submodel will contain an intercept. This results in a term corresponding to the subject term being included in the combined model (V10.V11.V12 in this instance).
The above model expands out further to: Each term in the expanded model corresponds to a variance component, so in this case, g=8.
When constructing Z all contrast information specified when the submodels are constructed in calls to g22yac is ignored and dummy variables are used throughout.
It is possible to store the cross-product matrix, ZTZ in a block diagonal form if Mr contains an overall subject effect, Sr. In this context Sr is defined as a main effect or interaction term that is contained in all other subject terms. For example, if the random effects model is constructed from 3 submodels with subject terms V1.V4, V1.V2.V4 and V1.V2.V3.V4, then Sr=V1.V4 and rnlsv will be set to the number of levels in Sr, otherwise rnlsv=1.

12 Optional Parameters

As well as the optional parameters common to all G22 handles described in g22zmc and g22znc, a number of additional optional parameters can be specified for a G22 handle holding the description of a linear mixed model, as returned by g02jfc in hlmm.
Each writeable optional parameter has an associated default value; to set any of them to a non-default value, use g22zmc. The value of any optional parameter can be queried using g22znc.
Most of the optional parameters described in this section are related to the behaviour g02jhc when fitting the model. These descriptions should, therefore, be read in conjunction with the documentation for that function.
The remainder of this section can be skipped if you wish to use the default values for all optional parameters.
The following is a list of the optional parameters available. A full description of each optional parameter is provided in Section 12.1.

12.1 Description of the Optional Parameters

For each option, we give a summary line, a description of the optional parameter and details of constraints.
The summary line contains:
Keywords and character values are case and white space insensitive.
Gamma Lower BoundrDefault =machine precision/100
A lower bound for the elements of γ*, where γ*=γ/σR2.
Gamma Upper BoundrDefault =1020
An upper bound for the elements of γ*, where γ*=γ/σR2.
Initial DistancerDefault =100000.0
The initial distance from the solution.
Initial Value StrategyiDefault =special
Controls how g02jhc will choose the initial values for the variance components, γ, if not supplied.
Initial Value Strategy=0
The MIVQUE0 estimates of the variance components based on the likelihood specified by Likelihood are used.
Initial Value Strategy=1
The MIVQUE0 estimates based on the maximum likelihood are used, irrespective of the value of Likelihood.
See Rao (1972) for a description of the minimum variance quadratic unbiased estimators (MIVQUE0).
By default, for small problems, Initial Value Strategy=0 and for large problems Initial Value Strategy=1.
Constraint: Initial Value Strategy=0 or 1.
LikelihoodaDefault =REML
Likelihood defines whether g02jhc will use the restricted maximum likelihood (REML) or the maximum likelihood (ML) when fitting the model.
Constraint: Likelihood=REML or ML.
Linear Minimization AccuracyrDefault =0.9
The accuracy of the linear minimizations.
Line Search TolerancerDefault =0.9
The line search tolerance.
List
NoList Default
Optional parameter List enables printing of each optional parameter specification as it is supplied. NoList suppresses this printing.
Major Iteration LimitiDefault =special
The number of major iterations.
Major Print LeveliDefault =special
The frequency that monitoring information is output to Unit Number.
Maximum Number of ThreadsiDefault =special
Controls the maximum number of threads used by g02jhc in a multithreaded library. By default, the maximum number of available threads are used.
In a library that is not multithreaded, this option has no effect.
Constraint: Maximum Number of Threads0.
Minor Iteration LimitiDefault =max(50,3×g)
The number of minor iterations.
Minor Print LeveliDefault =0
The frequency that additional monitoring information is output to Unit Number.
Optimality TolerancerDefault =machine precision0.72
The optimality tolerance.
Parallelisation StrategyiDefault =special
If Maximum Number of Threads>0 then Parallelisation Strategy controls how g02jhc is parallelised in a multithreaded library.
Parallelisation Strategy=1
g02jhc will attempt to parallelise operations involving Z, even if rnlsv=1.
Parallelisation Strategy=2
g02jhc will only attempt to parallelise operations involving Z, if rnlsv>1.
By default, Parallelisation Strategy=1, however, for some models / datasets, this may be slower than using Parallelisation Strategy=2 when rnlsv=1.
In a library that is not multithreaded, this option has no effect.
Constraint: Parallelisation Strategy=1 or 2.
Solution AccuracyrDefault =0.0
The accuracy to which the solution is required.
SolveraDefault =special
Controls which solver g02jhc will use when fitting the model. By default, Solver=E04LB is used for small problems and Solver=E04UC, otherwise.
If Solver=E04LB, then the solver used is the one implemented in e04lbc and if Solver=E04UC, then the solver used is the one implemented in e04ucc.
Constraint: Solver=E04LB or E04UC.
Sweep TolerancerDefault =special
The sweep tolerance used by g02jhc when performing the sweep operation Wolfinger et al. (1994). The default value used is Sweep Tolerance=max(ε,ε×(maxi(ZT)ii)), where ε=machine precision.
Unit NumberiDefault =Nag_FileID number associated with stdout
The monitoring Nag_FileID number (as returned from x04acc, stdout as the default) to which g02jhc will send any monitoring information.