Give Me 30 Minutes And I’ll Give You Autocorrelation
This can be frustrating because if you try to do a regression analysis on data with autocorrelation, then your analysis will be misleading. In most cases, this approach should work. The null hypothesis and alternate hypothesis for the Durbin-Watson Test areThe Durbin Watson test has values between 0 and 4. The following chart shows a random pattern, suggesting no autocorrelation:You can also make a correlogram [7], which is sometimes combined with a measure of correlation like Morans I. Fig 2. 8.
3 Things That Will Trip You Up In An Sari-Bradley Tests
However, you also have lag-2, lag-3, etc.
For example, if
X
=
(
X
1
X
2
,
X
3
)
T
{\displaystyle \mathbf {X} =\left(X_{1},X_{2},X_{3}\right)^{\rm {T}}}
is a random vector, then
R
X
link
X
{\displaystyle \operatorname {R} _{\mathbf {X} \mathbf {X} }}
is a
3
3
{\displaystyle 3\times 3}
matrix whose
(
i
,
j
)
{\displaystyle (i,j)}
-th entry is
E
[
X
i
X
j
]
{\displaystyle \operatorname {E} [X_{i}X_{j}]}
. .