This paper begins with the theory of identification of non Gaussian process in additive colored Gaussian noise, then analyses and reviews the cumulant methods to identify non Gaussian and non minimum phase ARMA model, which have been developed during recent years.

Gaussian process and its application to soft-sensor modeling;

Time series prediction of foundation pit displacement using Gaussian process method;

The law of iterated logarithm with finite partial sum for the Gaussian process;

We mainly use the discrete stationary Gauss processes as the input and output signals of two-dimensional linear system to estimate the impulse transfer function,and we discuss some properties about the estimation of the impulse transfer function.

With Gauss process and EI method calibration individuals can be selected effectively.

Gaussian processes(GP) are probabilistic kernel machines and moderately simple to implement and use without loss of performance compared with other kernel methods.

The asymptotic distributions of M(α) uniform for α∈S p-1  are thereby derived, which are supremums of the Gaussian processes on S p-1 .

We propose a novel modeling approach using Gaussian processes(GP).