Quiz 3
Registry Synced

Goodness of Fit

83 words
1 min read

Reading compass

Now · R-squared

Goodness of Fit

R-squared

R2=1−SSresSStot=SSregSStotR^2 = 1 - \frac{SS_{res}}{SS_{tot}} = \frac{SS_{reg}}{SS_{tot}}
  • Proportion of variance explained by the model
  • Always increases when adding predictors

Adjusted R-squared

Radj2=1−SSres/(n−p)SStot/(n−1)R^2_{adj} = 1 - \frac{SS_{res}/(n-p)}{SS_{tot}/(n-1)}
  • Penalizes model complexity
  • Can decrease when adding useless predictors

Partial F-Test

Test whether a subset of predictors collectively contribute:
F=(SSres,R−SSres,F)/(dfR−dfF)SSres,F/dfF∼FdfR−dfF,dfFF = \frac{(SS_{res,R} - SS_{res,F}) / (df_R - df_F)}{SS_{res,F} / df_F} \sim F_{df_R-df_F, df_F}
Join Discord PreviousMultiple Linear RegressionNextGauss-Markov Theorem
Document outline

Keep your place and jump directly to a heading.

Table of contents

No note selected

Open a chapter to see its outline, backlinks, and sources here.

Open libraryCheckpoint desk