help canon dialog: canon
also see: canon postestimation
-------------------------------------------------------------------------------
Title
[MV] canon -- Canonical correlations
Syntax
canon (varlist1) (varlist2) [if] [in] [weight] [, options]
options description
-------------------------------------------------------------------------
Model
lc(#) calculate the linear combinations for canonical
correlation #
first(#) calculate the linear combinations for the first #
canonical correlations
noconstant do not subtract means when calculating correlations
Reporting
stdcoef output matrices of standardized coefficients
stderr display raw coefficients and conditionally estimated
standard errors
level(#) set confidence level; default is level(95)
test(numlist) display significance tests for the specified canonical
correlations
notests do not display tests
format(%fmt) numerical format for coefficient matrices; default is
format(%8.4f)
-------------------------------------------------------------------------
by and statsby are allowed; see prefix.
aweights and fweights are allowed; see weight.
See [MV] canon postestimation for features available after estimation.
Menu
Statistics > Multivariate analysis > MANOVA, multivariate regression, and
related > Canonical correlations
Description
canon estimates canonical correlations and provides the coefficients for
calculating the appropriate linear combinations corresponding to those
correlations.
canon typed without arguments redisplays previous estimation results.
Options
+-------+
----+ Model +------------------------------------------------------------
lc(#) specifies that linear combinations for canonical correlation # be
calculated. By default, all are calculated.
first(#) specifies that linear combinations for the first # canonical
correlations be calculated. By default, all are calculated.
noconstant specifies that means not be subtracted when calculating
correlations.
+-----------+
----+ Reporting +--------------------------------------------------------
stdcoef specifies that the first part of the output contain the standard
coefficients of the canonical correlations in matrix form. The
default is to present the raw coefficients of the canonical
correlations in matrix form.
stderr specifies that the first part of the output contains the raw
coefficients of the canonical correlations, the conditionally
estimated standard errors, and the conditionally estimated confidence
intervals in the standard estimation table. The default is to
present the raw coefficients of the canonical correlations in matrix
form.
level(#) specifies the confidence level, as a percentage, for confidence
intervals of the coefficients. These "confidence intervals" are the
result of an approximate calculation; see [MV] canon.
test(numlist) specifies that significance tests of the canonical
correlations in the numlist be displayed. Because of the nature of
significance testing, if there are three canonical correlations,
test(1) will test the significance of all three correlations, test(2)
will test the significance of canonical correlations 2 and 3, and
test(3) will test the significance of the third canonical correlation
alone.
notests specifies that significance tests of the canonical correlations
not be displayed.
format(%fmt) specifies the display format for numbers in coefficient
matrices. format(%8.4f) is the default. format() may not be
specified with stderr.
Examples
Setup
. sysuse auto
Estimate canonical correlations
. canon (length weight headroom trunk) (displ mpg gear_ratio turn)
Same as above, but display standardized coefficients
. canon (length weight headroom trunk) (displ mpg gear_ratio turn),
stdcoef
Display raw coefficients and conditional standard errors in the standard
estimation table
. canon (length weight headroom trunk) (displ mpg gear_ratio turn),
stderr
Saved results
canon saves the following in e():
Scalars
e(N) number of observations
e(df_r) residual degrees of freedom
e(df) degrees of freedom
e(df1) numerator degrees of freedom for significance
tests
e(df2) denominator degrees of freedom for
significance tests
e(n_lc) the linear combination calculated
e(n_cc) number of canonical correlations calculated
e(rank) rank of e(V)
Macros
e(cmd) canon
e(cmdline) command as typed
e(wtype) weight type
e(wexp) weight expression
e(properties) b V
e(estat_cmd) program used to implement estat
e(predict) program used to implement predict
e(marginsnotok) predictions disallowed by margins
Matrices
e(b) coefficient vector
e(stat_#) statistics for canonical correlation #
e(stat_m) statistics for overall model
e(canload11) canonical loadings for varlist_1
e(canload22) canonical loadings for varlist_2
e(canload12) correlation between varlist_1 and the
canonical variates from varlist_2
e(canload21) correlation between varlist_2 and the
canonical variates from varlist_1
e(rawcoef_var1) raw coefficients for varlist_1
e(rawcoef_var2) raw coefficients for varlist_2
e(stdcoef_var1) standard coefficients for varlist_1
e(stdcoef_var2) standard coefficients for varlist_2
e(ccorr) canonical correlation coefficients
e(ccorr_var1) correlation matrix for varlist_1
e(ccorr_var2) correlation matrix for varlist_2
e(ccorr_mixed) correlation matrix between varlist_1 and
varlist_2
e(V) variance-covariance matrix of the estimators
Functions
e(sample) marks estimation sample
Also see
Manual: [MV] canon
Help: [MV] canon postestimation;
[MV] factor, [MV] pca, [R] correlate, [R] mvreg, [R] pcorr, [R]
regress
