Way of inferring information from cross-covariance matrices
In
statistics, canonical-correlation analysis (CCA), also called canonical variates analysis, is a way of inferring information from
cross-covariance matrices. If we have two vectors X = (X1, ..., Xn) and Y = (Y1, ..., Ym) of
random variables, and there are
correlations among the variables, then canonical-correlation analysis will find linear combinations of X and Y that have a maximum correlation with each other.[1] T. R. Knapp notes that "virtually all of the commonly encountered
parametric tests of significance can be treated as special cases of canonical-correlation analysis, which is the general procedure for investigating the relationships between two sets of variables."[2] The method was first introduced by
Harold Hotelling in 1936,[3] although in the context of
angles between flats the mathematical concept was published by
Camille Jordan in 1875.[4]
CCA is now a cornerstone of multivariate statistics and multi-view learning, and a great number of interpretations and extensions have been proposed, such as probabilistic CCA, sparse CCA, multi-view CCA, Deep CCA, and DeepGeoCCA.[5] Unfortunately, perhaps because of its popularity, the literature can be inconsistent with notation, we attempt to highlight such inconsistencies in this article to help the reader make best use of the existing literature and techniques available.
Like its sister method
PCA, CCA can be viewed in population form (corresponding to random vectors and their covariance matrices) or in sample form (corresponding to datasets and their sample covariance matrices). These two forms are almost exact analogues of each other, which is why their distinction is often overlooked, but they can behave very differently in high dimensional settings.[6] We next give explicit mathematical definitions for the population problem and highlight the different objects in the so-called canonical decomposition - understanding the differences between this objects is crucial for interpretation of the technique.
Canonical-correlation analysis seeks a sequence of vectors () and () such that the random variables and maximize the
correlation. The (scalar) random variables and are the first pair of canonical variables. Then one seeks vectors maximizing the same correlation subject to the constraint that they are to be uncorrelated with the first pair of canonical variables; this gives the second pair of canonical variables. This procedure may be continued up to times.
The sets of vectors are called canonical directions or weight vectors or simply weights. The 'dual' sets of vectors are called canonical loading vectors or simply loadings; these are often more straightforward to interpret than the weights.[7]
Computation
Derivation
Let be the
cross-covariance matrix for any pair of (vector-shaped) random variables and . The target function to maximize is
There is equality if the vectors and are collinear. In addition, the maximum of correlation is attained if is the
eigenvector with the maximum eigenvalue for the matrix (see
Rayleigh quotient). The subsequent pairs are found by using
eigenvalues of decreasing magnitudes. Orthogonality is guaranteed by the symmetry of the correlation matrices.
Another way of viewing this computation is that and are the left and right
singular vectors of the correlation matrix of X and Y corresponding to the highest singular value.
R as the standard function
cancor and several other packages, including
CCA and
vegan.
CCP for statistical hypothesis testing in canonical correlation analysis.
Each row can be tested for significance with the following method. Since the correlations are sorted, saying that row is zero implies all further correlations are also zero. If we have independent observations in a sample and is the estimated correlation for . For the th row, the test statistic is:
which is asymptotically distributed as a
chi-squared with degrees of freedom for large .[12] Since all the correlations from to are logically zero (and estimated that way also) the product for the terms after this point is irrelevant.
Note that in the small sample size limit with then we are guaranteed that the top correlations will be identically 1 and hence the test is meaningless.[13]
Practical uses
A typical use for canonical correlation in the experimental context is to take two sets of variables and see what is common among the two sets.[14] For example, in psychological testing, one could take two well established multidimensional
personality tests such as the
Minnesota Multiphasic Personality Inventory (MMPI-2) and the
NEO. By seeing how the MMPI-2 factors relate to the NEO factors, one could gain insight into what dimensions were common between the tests and how much variance was shared. For example, one might find that an
extraversion or
neuroticism dimension accounted for a substantial amount of shared variance between the two tests.
One can also use canonical-correlation analysis to produce a model equation which relates two sets of variables, for example a set of performance measures and a set of explanatory variables, or a set of outputs and set of inputs. Constraint restrictions can be imposed on such a model to ensure it reflects theoretical requirements or intuitively obvious conditions. This type of model is known as a maximum correlation model.[15]
Visualization of the results of canonical correlation is usually through bar plots of the coefficients of the two sets of variables for the pairs of canonical variates showing significant correlation. Some authors suggest that they are best visualized by plotting them as heliographs, a circular format with ray like bars, with each half representing the two sets of variables.[16]
The definition of the canonical variables and is then equivalent to the definition of
principal vectors for the pair of subspaces spanned by the entries of and with respect to this
inner product. The canonical correlations is equal to the
cosine of
principal angles.
Whitening and probabilistic canonical correlation analysis
CCA can also be viewed as a special
whitening transformation where the random vectors and are simultaneously transformed in such a way that the cross-correlation between the whitened vectors and is diagonal.[17]
The canonical correlations are then interpreted as regression coefficients linking and and may also be negative. The regression view of CCA also provides a way to construct a latent variable probabilistic generative model for CCA, with uncorrelated hidden variables representing shared and non-shared variability.
^Knapp, T. R. (1978). "Canonical correlation analysis: A general parametric significance-testing system". Psychological Bulletin. 85 (2): 410–416.
doi:
10.1037/0033-2909.85.2.410.
^Yang Song, Peter J. Schreier, David Ram´ırez, and Tanuj Hasija Canonical correlation analysis of high-dimensional data with very small sample supportarXiv:
1604.02047
^Tofallis, C. (1999). "Model Building with Multiple Dependent Variables and Constraints". Journal of the Royal Statistical Society, Series D. 48 (3): 371–378.
arXiv:1109.0725.
doi:
10.1111/1467-9884.00195.
S2CID8942357.