Chunk #74 — Results — An Estimate for the Data Size Needed for Significance
Text
Let l 1 be the lead eigenvalue of the theoretical covariance matrix, with the remainder of the eigenvalues 1. Set γ 2 = n/m. Let L 1 be the largest eigenvalue of the sample covariance. We will let n, m become large with the ratio n/m tending to a limit.