“…After identifying the dimension where it has most variation, it is possible to determine the best approximation of the original data with fewer dimensions. The SVD [11,12] of a matrix X m×n is given as X = UΣ V T where U ∈ m×m , V ∈ n×n , and Σ m×n = [diag{σ 1 , · · · , σ r }:0], r = min(m, n), and σ 1 , σ 2 , · · · , σ r are the singular values. The reduced SVD of the data matrix can be given as X =Û m×nΣ n×n V T n×n where n and m represent the number of leads and corresponding samples of each lead, respectively [11].…”