These methods use a suitable set of canonical variables obtained from the covariance matrices of the training and probe image pairs in the two spaces to identify the subject [12].
The two correlation matrices of the training and control data sets developed as R X = [ r ij ] and R Y = [ r ij ] and revealed the pattern of TFs existing in the two comparative data sets (Additional file 7: Table S6).
The Mahalanobis distances d0, u and d0, t are measured in the model space defined by b, i.e., they are weighted Euclidean distances between the centroid X0 of the descriptor matrix for the training set and the vectors Xu and Xt, respectively.
According to the aggregation factors, the rather diffuse neighbourhood matrix of the unsupervised training becomes sharper by supervising the training, which amplifies the correlation between closer CWT's.
Let the diagonal matrix XS=diag xS consist of the training symbols at the source node; then the received signals in (1) are further written in a matrix form as z SD = X S Φ V SD h SD + n t (2).
Build training matrix: To avoid duplicate calculating the frequency of a test and the joint frequency between two tests, we build a training matrix out of the training data.
Using Equation 12, the mean μ y Te and covariance Σ y Te of y Te can be expressed from the mean vector and covariance matrix of the noisy training speech y Tr as follows: μ y Te = μ y Tr + G ( μ y Tr, n 0, n Tr ) + ∇ n G ( μ y Tr, n 0, n Tr ) ( n - n 0 ) Σ y Te = I + ∇ y Tr G μ y Tr, n 0, n Tr Σ y Tr I + ∇ y Tr G μ y Tr, n 0, n Tr T, (14).
Second, that the smoothed matrix of the 1992 93 training data represents better the 1994 test data than the unsmoothed one as it gives better prediction, which was also confirmed for the periods 1975 94/1995.
