ABSTRACT
Most signals in practice are not deterministic but random. However, they can be described by precise mathematical analysis whose tools are contained in the theory statistical analysis. A discrete process is purely random if the random variables of a sequence are mutually independent and identically distributed variables. A Gaussian random process has the following properties: it is completely defined by its mean vector and covariance matrix; any linear operation on the time variables produces another Gaussian random process; all higher moments can be expressed by the first and second moments of the distribution; and white noise is necessarily generated by independent, identically distributed samples. Because the correlation function at its extreme lag values is not reliable due to the small overlapping of the correlation process, it is recommended to use lag values of about 30 to 40% of the total length of the data.
