The tool supports interface of Foresight Analytics Platform 9 or earlier.
The test checks residuals of a linear regression model for heteroscedasticity.
Test parameters:
Explanatory Variables. Factors that affect the behavior of the output variable. By default, the list includes all factors of the tested linear regression model. The Factor checkbox indicates that the factor is included into the test. All factors are involved in the test by default. Deselect the checkbox to exclude the factor from the test. The number of explanatory variables should be at least one.
Cross Products. Selecting the checkbox enables the use of both the existing explanatory series and their cross products. For example, the S1 and the S2 series are considered. Using the cross products, the following regressors are considered: S1*S1, S1*S2, S2*S2. It is not recommended to include cross products if the number of explanatory variables (regressors) is large.
Significance Level. The significance level value, at which the hypothesis is rejected.
The results are displayed as a table containing the following data:
The following is given for each statistics: value, statistics probability and the test result (whether the hypothesis on homoscedasticity of residuals is accepted or rejected).
Model Coefficients. Regression coefficients calculated for the selected factors (including cross products if they are used).
NOTE. If test parameters have been specified incorrectly, the result table is not displayed. An error message is displayed instead of the table.
The example of the results table:
See also: