The multivariate normal distribution as a generalization of the familiar univariate normal or Gaussian distribution to p ≥ 2 dimensions. Just as with its univariate counterpart, the importance of the multivariate normal distribution emanates from a number of its useful properties, and especially from the fact that, according to the central limit theorem under certain regularity conditions, sums of random variables generated from various (likely unknown) distributions tend to behave as if its underlying distribution were multivariate normal.