This chapter focuses on two random variables. The random variable is a useful concept for evaluating the parameters for variables such as mean and standard deviation. In engineering, many interesting applications, such as autocorrelation, cross-correlation, and covariance, can be handled by the theory of two random variables. The commonly used models for signals can be classified as parametric or nonparametric models. The methods that do not make any assumption about how the data are generated in a signal are called nonparametric methods. Nonparametric models use statistical parameters for signal modeling. Parametric methods rely on the process of signal generation. The autoregressive (AR) model is commonly used model for signals like speech. AR systems have only poles, and zeros exist only at zero. Parametric methods are based on the modeling of data sequence as the output of a linear system characterized by the transfer function with poles and zeros.