Checking the degree of sample dependence
In statistical analysis, it is important to understand whether we are dealing with dependent or independent sets of data. This affects the system model building approach, and thus, the forecast quality.
The independence of the two data samples (represented in a vector or matrix form) is carried out with the help of the IndependenceTest function. This function conducts a series of tests to check the main hypothesis H0 to see whether the vectors are independent, and to check the alternative hypothesis HA to see whether the vectors are dependent.
The following tests are conducted:
|
Test's name |
Type of test |
Description |
|---|---|---|
|
"BlomqvistBeta" |
Monotonic |
This is based on Blomqvist's β |
|
"GoodmanKruskalGamma" |
Monotonic, vector |
This is based on the γ-coefficient |
|
"HoeffdingD" |
Vector |
This is based on Hoeffding's D |
|
"KendallTau" |
Monotonic |
This is based on Kendall's τ-b |
|
"PearsonCorrelation" |
Linear, normality, vector |
This is based on Pearson's product-moment... |
