2 problems

Let the two data streams be those used by the sliding-window canonical correlation analysis method, and suppose that the data satisfy relatively mild conditions, for example, that…

Let $d$ be the dimension, let $n$ be a fixed sample size, and let $\alpha$ be the population eigenvalue decay parameter. Denote the first sample canonical weight vector by…