lack_of_fit: Conduct the F Test for Lack of Fit

In ryan-heslin/RegLesson: Linear Regression from Scratch

Conduct the F Test for Lack of Fit

Description

Conduct the F Test for Lack of Fit

Usage

lack_of_fit(
  X,
  Y,
  Y_hat,
  resid = sum_square(Y, Y_hat),
  n = NROW(Y),
  p = NCOL(X),
  alpha = 0.05
)

Arguments

X	An n \times m design matrix, optionally with an intercept column present.
Y	An n-length vector or n \times 1 matrix representing the response.
Y_hat	Vector of fitted values
resid	Vector of residuals
n	Number of rows in the design matrix. Computed if not provided.
p	Number of columns in the design matrix. Computed if not provided.
alpha	Alpha level for the significance test.

Details

This function carries out the F test for lack of fit, which essentially compares the model estimates to the local means of the response at each distinct observation. The test requires duplicate observations.

Value

Computed F statistic for the test.

ryan-heslin/RegLesson documentation built on Aug. 5, 2022, 9:03 p.m.