The function may be called by the names: g03acc or nag_mv_canon_var.
Let a sample of observations on variables in a data matrix come from groups with observations in each group, . Canonical variate analysis finds the linear combination of the variables that maximizes the ratio of between-group to within-group variation. The variables formed, the canonical variates can then be used to discriminate between groups.
The canonical variates can be calculated from the eigenvectors of the within-group sums of squares and cross-products matrix. However, g03acc calculates the canonical variates by means of a singular value decomposition (SVD) of a matrix . Let the data matrix with variable (column) means subtracted be , and let its rank be ; then the matrix is given by:
where is an orthogonal matrix that defines the groups and is the first rows of the orthogonal matrix either from the decomposition of :
if is of full column rank, i.e., , else from the SVD of :
Let the SVD of be:
then the nonzero elements of the diagonal matrix , , for , are the canonical correlations associated with the canonical variates, where .
The eigenvalues, , of the within-group sums of squares matrix are given by:
and the value of gives the proportion of variation explained by the th canonical variate. The values of the 's give an indication as to how many canonical variates are needed to adequately describe the data, i.e., the dimensionality of the problem.
To test for a significant dimensionality greater than the statistic:
can be used. This is asymptotically distributed as a distribution with degrees of freedom. If the test for is not significant, then the remaining tests for should be ignored.
The loadings for the canonical variates are calculated from the matrix . This matrix is scaled so that the canonical variates have unit within group variance.
In addition to the canonical variates loadings the means for each canonical variate are calculated for each group.
Weights can be used with the analysis, in which case the weighted means are subtracted from each column and then each row is scaled by an amount , where is the weight for the th observation (row).
Chatfield C and Collins A J (1980) Introduction to Multivariate Analysis Chapman and Hall
Gnanadesikan R (1977) Methods for Statistical Data Analysis of Multivariate Observations Wiley
Hammarling S (1985) The singular value decomposition in multivariate statistics SIGNUM Newsl.20(3) 2–25
Kendall M G and Stuart A (1979) The Advanced Theory of Statistics (3 Volumes) (4th Edition) Griffin
1: – Nag_WeightstypeInput
On entry: indicates the type of weights to be used in the analysis.
No weights are used.
The weights are treated as frequencies and the effective number of observations is the sum of the weights.
The weights are treated as being inversely proportional to the variance of the observations and the effective number of observations is the number of observations with nonzero weights.
, or .
2: – IntegerInput
On entry: the number of observations, .
3: – IntegerInput
On entry: the total number of variables, .
4: – const doubleInput
On entry: must contain the th observation for the th variable, for and .
5: – IntegerInput
On entry: the stride separating matrix column elements in the array x.
6: – const IntegerInput
On entry: indicates whether or not the th variable is to be included in the analysis.
If , then the variable contained in the th column of x is included in the canonical variate analysis, for .
On entry: the number of variables in the analysis, .
8: – const IntegerInput
On entry: indicates which group the th observation is in, for . The effective number of groups is the number of groups with nonzero membership.
, for .
9: – IntegerInput
On entry: the number of groups, .
10: – const doubleInput
On entry: if or then the elements of wt must contain the weights to be used in the analysis.
If then the th observation is not included in the analysis.
, for ;
effective number of groups.
Note: if then wt is not referenced and may be NULL.
11: – IntegerOutput
On exit: gives the number of observations in group , for .
12: – doubleOutput
On exit: contains the mean of the th canonical variate for the th group, for and ; the remaining columns, if any, are used as workspace.
13: – IntegerInput
On entry: the stride separating matrix column elements in the array cvm.
14: – doubleOutput
On exit: the statistics of the canonical variate analysis. , the canonical correlations, , for .
, the eigenvalues of the within-group sum of squares matrix, , for .
, the proportion of variation explained by the th canonical variate, for .
, the statistic for the th canonical variate, for .
, the degrees of freedom for statistic for the th canonical variate, for .
, the significance level for the statistic for the th canonical variate, for .
15: – IntegerInput
On entry: the stride separating matrix column elements in the array e.
16: – Integer *Output
On exit: the number of canonical variates, . This will be the minimum of and the rank of x.
17: – doubleOutput
On exit: the canonical variate loadings. contains the loading coefficient for the th variable on the th canonical variate, for and ; the remaining columns, if any, are used as workspace.
18: – IntegerInput
On entry: the stride separating matrix column elements in the array cvx.
19: – doubleInput
On entry: the value of tol is used to decide if the variables are of full rank and, if not, what is the rank of the variables. The smaller the value of tol the stricter the criterion for selecting the singular value decomposition. If a non-negative value of tol less than machine precision is entered, then the square root of machine precision is used instead.
20: – Integer *Output
On exit: the rank of the dependent variables.
If the variables are of full rank then .
If the variables are not of full rank then irankx is an estimate of the rank of the dependent variables. irankx is calculated as the number of singular values greater than (largest singular value).
21: – NagError *Input/Output
The NAG error argument (see Section 7 in the Introduction to the NAG Library CL Interface).
The singular value decomposition has failed to converge. This is an unlikely error exit.
The number of variables, nx in the analysis , while number of variables included in the analysis via array .
Constraint: these two numbers must be the same.
The wt array argument must not be NULL when the weight argument indicates weights.
As the computation involves the use of orthogonal matrices and a singular value decomposition rather than the traditional computing of a sum of squares matrix and the use of an eigenvalue decomposition, g03acc should be less affected by ill conditioned problems.
8Parallelism and Performance
Background information to multithreading can be found in the Multithreading documentation.
g03acc is not threaded in any implementation.
A sample of nine observations, each consisting of three variables plus group indicator, is read in. There are three groups. An unweighted canonical variate analysis is performed and the results printed.