Moving AnalysisSingle Self Correlation Analysis studies the relationship between the value and the value of previous 1 step. This is the first step of Self Correlation Analysis.
AR Model is more general formulation of self correlation.
AR stands for "Auto Regressive."
If each step is affected from some former steps, AR model is useful. AR model is one of the Multi Regression Analysis .
If we consider the variation of AR model, there are many patterns. Examples are below. By the change and fusion of ideas of formulation, there are various patterns.
This is the more general formulation of SCA.
If single correlation between x(n) and x(n-1) is strong, the formulation could be
x(n) = A * (x-1).
If single correlation between x(n) and x(n-1) is not strong,
we may find suitable formulation between x(n) and x(n-1) by the scatter plot of x(n) and x(n-1).
x(n) = x(n-1)
means that the value does not change.
This is the application of (1). It is a general formulation of AR model using former 2 steps.
For example,
x(n) = x(n-1) + 0.01 x(n-2)
means that the former 2 step effects by 1%.
The time of both sides of the formulation is the same.
So it seems that this formulation is not for the time series analysis. But it is used in Condition Analysis. And this formulation is used if the measurement of x(n) needs much time to get.
(3) is needed to understand (4) and (5) more easily.
u and v means causes. And x means an effect.
This is same to
x(n) - x(n-1) = f[y(n-1),z(n-1)].
.
It means that "difference of x is caused by other variables."
This formulation means that "The value of former t hours repeats." It means periodicity.
We can use (6) for Spectrum Analysis.
Analysis for non-linear data is studied in chaos field.
NEXT Moving Analysis
