Skip to content

vlt.stats.nestedftest

  NESTEDFTEST - Compute Nested F test statistic for 2 nested models

   [P,F] = vlt.stats.nestedftest(SE1, P1, SE2, P2, NDATA)

   Performs a Nested F test for 2 model fits to the same data, where
   the first model is "nested" inside the second (that is, the first model
   is a simplified version of the second, and can be obtained from the
   second model by setting parameters in the second to particular values).

   SE1: The total squared error between the first model and the data.
   P1:  The number of parameters in the first model.
   SE2: The total squared error between the second model and the data.
   P2:  The number of parameters in the second model.
   NDATA: The number of data values used to calculate SE1, SE2.

   P:   The p-value of the F test.  If this is less than your alpha (say,
        0.05) then the second model is a significantly better fit than the
        first.
   F:   The value of the F statistic.

   See: http://en.wikipedia.org/wiki/F-test#Regression_problems