Merge pull request #53 from nalimilan/patch-1

Fix reference to saturated model
yufongpeng · Sep 12, 2022 · 0ebdb71 · 0ebdb71
2 parents 657919b + ecff455
commit 0ebdb71
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/docs/src/GLM.md b/docs/src/GLM.md
@@ -43,7 +43,7 @@ This result is a little bit different from [`GLM.ftest`](https://juliastats.org/
 ```@example glm
 ftest(getproperty.(lms[2:end], :model)...)
 ```
-In `anova`, the F value is calculated by dividing [MSR](https://en.wikipedia.org/wiki/Mean_squared_error) (mean of ΔDeviance) with mean of [RSS](https://en.wikipedia.org/wiki/Residual_sum_of_squares) of saturated model just like [`anova` in R](https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/anova), while in [`GLM.ftest`](https://juliastats.org/GLM.jl/stable/api/#GLM.ftest), the denominator is replaced by [RSS](https://en.wikipedia.org/wiki/Residual_sum_of_squares) of subsequant model.
+In `anova`, the F value is calculated by dividing [MSR](https://en.wikipedia.org/wiki/Mean_squared_error) (mean of ΔDeviance) with mean of [RSS](https://en.wikipedia.org/wiki/Residual_sum_of_squares) of the most complex model just like [`anova` in R](https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/anova), while in [`GLM.ftest`](https://juliastats.org/GLM.jl/stable/api/#GLM.ftest), the denominator is replaced by [RSS](https://en.wikipedia.org/wiki/Residual_sum_of_squares) of subsequent model.
 ## Generalized linear models 
 ```julia
 quine = dataset("MASS", "quine")
@@ -72,4 +72,4 @@ lrtest(glm1, glm2)
 ```
 This function works identically as [`StatsModels.lrtest`](https://juliastats.org/StatsModels.jl/stable/api/#StatsModels.lrtest).
 !!! note
-    We can also specify test by keword arguments `test` or putting test in the first argument.
+    We can also specify test by keword arguments `test` or putting test in the first argument.