gbxcorr1

Usage

Overview

Take as input a data matrix A with N rows and T columns

A = [ a_{t,i} ]

with t=1,..,T i=1,...,N

and compute the NxN correlation matrix C [ c_{i,j} ] following the method specified by option '-M': with method 0 it is

c_{i,j} = 1/(T-1) \sum_t (a_{t,i}-m_i) (a_{t,j}-m_j)

where m_i is the i-th mean and with method 1 it is

c_{i,j} = 1/T \sum_t a_{t,i} a_{t,j} .

Covariance is stored in the lower triangle while correlation coefficients are stored in the upper triangle.

WARNING: previous implementations were row-wise instead of column-wise

Options

-M

choose the method (default 0)

-F specify the input fields separators (default " \t")

-h this help