Using a Table of Coefficients


The table of coefficients provides the model coefficients for the parametric models (linear and logistic regression). In addition to the estimates of the coefficients, the table includes a measure of the variability or error of each estimate and a test statistic (t.value or z.value) of the null hypothesis that the coefficients is zero (in other words, not needed in the model). A p-value for the statistical test is also provided. A small p-value (typically less that 0.05) indicates that the null hypothesis can be rejected--that is, that the coefficient is significant or important in the model.

Note that the test for each coefficient is based on the model with all the other coefficients included in the model.

The absolute value of the test statistic can provide a measure of variable importance for that term in the model. The larger the absolute value, the greater the importance.

The table of coefficients is added as a separate data table in Spotfire and can be used for further analysis.

See also:

Using a Model Summary

Available Diagnostic Visualizations