Name

CL_stat — Statistics on a matrix of samples

Calling Sequence

   [means,cov] = CL_stat(samples)
   
   

Description

  • Computes the mean and the covariance matrix of samples. samples is a numerical matrix PxN who gives the values of P variables for N individuals.
  • Last update : 10/11/2009

Parameters

samples:

matrix of sample values. It represents N samples of a vector of P variables. (PxN)

means:

vector of means (means(i) is the mean of samples(i,:)) (Px1)

cov:

covariance matrix (PxP)

Authors

CNES - DCT/SB

See also

CL_cor2cov

Examples

//given correlation in cartesian coordinates, compute correlation in adapted circular parameters with Monte Carlo process
bulletin = [-1877901 -3909428 -5026025 7428.157 -1541.857 -1576.532]'; //position and velocity
cor = [1 -0.467016 -0.447601 0.960396 0.987145 0.995826;...
0 1 -0.088751 -0.359696 -0.412472 -0.540655;...
0 0 1 -0.248472 -0.582834 -0.431908;...
0 0 0 1 0.915197 0.943178;...
0 0 0 0 1 0.980679;...
0 0 0 0 0 1]; //correlation matrix
cor = cor+cor'-eye(cor);  //complete correlation matrix (symmetric)
sd = [15939.68154 2912.099353 3079.494708 6.81910416 9.50017639 12.14624495]'; //standard deviations
cov = CL_cor2cov(cor,sd);
drawn_car = CL_covDraw(bulletin,cov,10000);  //draw 10000 samples following correlation
drawn_cir = CL_oe_car2cir(drawn_car(1:3,:),drawn_car(4:6,:));  //convert samples to adapted circular par.
[mean_cir,cov_cir] = CL_stat(drawn_cir);