g02de adds a new independent variable to a general linear regression model.

Syntax

C#
public static void g02de(
	string weight,
	int n,
	int ip,
	double[,] q,
	double[] p,
	double[] wt,
	double[] x,
	out double rss,
	double tol,
	out int ifail
)
Visual Basic
Public Shared Sub g02de ( _
	weight As String, _
	n As Integer, _
	ip As Integer, _
	q As Double(,), _
	p As Double(), _
	wt As Double(), _
	x As Double(), _
	<OutAttribute> ByRef rss As Double, _
	tol As Double, _
	<OutAttribute> ByRef ifail As Integer _
)
Visual C++
public:
static void g02de(
	String^ weight, 
	int n, 
	int ip, 
	array<double,2>^ q, 
	array<double>^ p, 
	array<double>^ wt, 
	array<double>^ x, 
	[OutAttribute] double% rss, 
	double tol, 
	[OutAttribute] int% ifail
)
F#
static member g02de : 
        weight : string * 
        n : int * 
        ip : int * 
        q : float[,] * 
        p : float[] * 
        wt : float[] * 
        x : float[] * 
        rss : float byref * 
        tol : float * 
        ifail : int byref -> unit 

Parameters

weight
Type: System..::..String
On entry: indicates if weights are to be used.
weight="U"
Least squares estimation is used.
weight="W"
Weighted least squares is used and weights must be supplied in array wt.
Constraint: weight="U" or "W".
n
Type: System..::..Int32
On entry: n, the number of observations.
Constraint: n1.
ip
Type: System..::..Int32
On entry: p, the number of independent variables already in the model.
Constraint: ip0 and ip<n.
q
Type: array<System..::..Double,2>[,](,)[,][,]
An array of size [dim1, ip+2]
Note: dim1 must satisfy the constraint: dim1n
On entry: if ip0, q must contain the results of the QR decomposition for the model with p parameters as returned by g02da or a previous call to g02de.
If ip=0, the first column of q should contain the n values of the dependent variable, y.
On exit: the results of the QR decomposition for the model with p+1 parameters:
  • the first column of q contains the updated value of c;
  • the columns 2 to ip+1 are unchanged;
  • the first ip+1 elements of column ip+2 contain the new column of R, while the remaining n-ip-1 elements contain details of the matrix Qp+1.
p
Type: array<System..::..Double>[]()[][]
An array of size [ip+1]
On entry: contains further details of the QR decomposition used. The first ip elements of p must contain the zeta values for the QR decomposition (see (F08AEF not in this release) for details).
The first ip elements of array p are provided by g02da or by previous calls to g02de.
On exit: the first ip elements of p are unchanged and the ip+1th element contains the zeta value for Qp+1.
wt
Type: array<System..::..Double>[]()[][]
An array of size [dim1]
Note: the dimension of the array wt must be at least n if weight="W", and at least 1 otherwise.
On entry: if weight="W" , wt must contain the weights to be used.
If wt[i-1]=0.0, the ith observation is not included in the model, in which case the effective number of observations is the number of observations with nonzero weights.
If weight="U", wt is not referenced and the effective number of observations is n.
Constraint: if weight="W", wt[i-1]0.0, for i=1,2,,n.
x
Type: array<System..::..Double>[]()[][]
An array of size [n]
On entry: x, the new independent variable.
rss
Type: System..::..Double%
On exit: the residual sum of squares for the new fitted model.
Note:  this will only be valid if the model is of full rank, see [Further Comments].
tol
Type: System..::..Double
On entry: the value of tol is used to decide if the new independent variable is linearly related to independent variables already included in the model. If the new variable is linearly related then c is not updated. The smaller the value of tol the stricter the criterion for deciding if there is a linear relationship.
Suggested value: tol=0.000001.
Constraint: tol>0.0.
ifail
Type: System..::..Int32%
On exit: ifail=0 unless the method detects an error or a warning has been flagged (see [Error Indicators and Warnings]).

Description

A linear regression model may be built up by adding new independent variables to an existing model. g02de updates the QR decomposition used in the computation of the linear regression model. The QR decomposition may come from g02da or a previous call to g02de. The general linear regression model is defined by
y=Xβ+ε,
where y is a vector of n observations on the dependent variable,
X is an n by p matrix of the independent variables of column rank k,
β is a vector of length p of unknown parameters,
and ε is a vector of length n of unknown random errors such that varε=Vσ2, where V is a known diagonal matrix.
If V=I, the identity matrix, then least squares estimation is used. If VI, then for a given weight matrix WV-1, weighted least squares estimation is used.
The least squares estimates, β^ of the parameters β minimize y-XβTy-Xβ while the weighted least squares estimates, minimize y-XβTWy-Xβ.
The parameter estimates may be found by computing a QR decomposition of X (or W12X in the weighted case), i.e.,
X=QR*or  W12X=QR*,
where R*=R0 and R is a p by p upper triangular matrix and Q is an n by n orthogonal matrix.
If R is of full rank, then β^ is the solution to
Rβ^=c1,
where c=QTy (or QTW12y) and c1 is the first p elements of c.
If R is not of full rank a solution is obtained by means of a singular value decomposition (SVD) of R.
To add a new independent variable, xp+1, R and c have to be updated. The matrix Qp+1 is found such that Qp+1TR:QTxp+1 (or Qp+1TR:QTW12xp+1) is upper triangular. The vector c is then updated by multiplying by Qp+1T.
The new independent variable is tested to see if it is linearly related to the existing independent variables by checking that at least one of the values QTxp+1i, for i=p+2,,n, is nonzero.
The new parameter estimates, β^, can then be obtained by a call to g02dd.
The method can be used with p=0, in which case R and c are initialized.

References

Draper N R and Smith H (1985) Applied Regression Analysis (2nd Edition) Wiley
Golub G H and Van Loan C F (1996) Matrix Computations (3rd Edition) Johns Hopkins University Press, Baltimore
Hammarling S (1985) The singular value decomposition in multivariate statistics SIGNUM Newsl. 20(3) 2–25
McCullagh P and Nelder J A (1983) Generalized Linear Models Chapman and Hall
Searle S R (1971) Linear Models Wiley

Error Indicators and Warnings

Note: g02de may return useful information for one or more of the following detected errors or warnings.
Errors or warnings detected by the method:
Some error messages may refer to parameters that are dropped from this interface (LDQ) In these cases, an error in another parameter has usually caused an incorrect value to be inferred.
ifail=1
On entry,n<1,
orip<0,
oripn,
ortol0.0,
orweight"U" or "W".
ifail=3
The new independent variable is a linear combination of existing variables. The ip+2th column of q will therefore be null.
ifail=-9000
An error occured, see message report.
ifail=-6000
Invalid Parameters value
ifail=-4000
Invalid dimension for array value
ifail=-8000
Negative dimension for array value
ifail=-6000
Invalid Parameters value

Accuracy

The accuracy is closely related to the accuracy of f08ag which should be consulted for further details.

Parallelism and Performance

None.

Further Comments

It should be noted that the residual sum of squares produced by g02de may not be correct if the model to which the new independent variable is added is not of full rank. In such a case g02dd should be used to calculate the residual sum of squares.

Example

See Also