Date: Jun 8, 2012 11:06 PM
Author: kumar vishwajeet
Subject: ekpirical histogram of a density function

I have 10000 realizations of a 50-dimensional vector X. i.e. I have X1, X2,...X10000, each of size 50x1. How to plot the empirical histogram of density function of X??
My understanding:
1. Find the mean of 10000 samples. We get a 50 dimensional vector X_mean.
2. Find the covariance of 10000 samples. We get 50x50 dimensional matrix X_cov.
3. Fit a Gaussian distribution with X_mean and X_cov.
4. Draw a 50-dimensional sample X_sample from the Gaussian distribution.
5. [f,x] = ecdf(X_sample).
6. ecdfhist(f,x)

Is this approach correct???