I have a one-dimensional matrix (that was extracted from the audio data) to process so the noise of the data could be eliminated. A paper told me that I have to use the Karhunen-Loeve Transformation (KLT). But, when I search for the fundamental of KLT, it involves eigenvectors and eigenvalues of the matrix's covariance. I have 1 x n or n x 1 matrix so the covariance is a certain value not a matrix (or 1 x 1 matrix). Then, the eigenvectors must be one. But the KLT transform is: y = eigenvector(covariance(data))*data.. It is not changing anything. Is one-dimensional matrix can be transformed by KLT?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.