summary.estimate_slopes() no longer working.

[strengejacke]

Fixes #345

[strengejacke]

I see, from the gam-example, and now understand the "segments" idea. Ignore this for now, let's try to make the summary() like it was intended to be, not like I suggested here.

[strengejacke]

More like this?

library(modelbased)
model <- mgcv::gam(Sepal.Width ~ s(Petal.Length), data = iris)
slopes <- estimate_slopes(model, by = "Petal.Length", length = 50)
#> No numeric variable was specified for slope estimation. Selecting `trend
#>   = "Petal.Length"`.
summary(slopes)
#> Average Marginal Effects
#> 
#> Start |  End | Direction | Confidence     
#> ------------------------------------------
#> 1.00  | 1.60 | positive  | Not Significant
#> 1.72  | 1.96 | negative  | Not Significant
#> 2.08  | 3.05 | negative  | Significant    
#> 3.17  | 3.41 | negative  | Not Significant
#> 3.53  | 3.65 | positive  | Not Significant
#> 3.77  | 4.25 | positive  | Significant    
#> 4.37  | 6.18 | positive  | Not Significant
#> 6.30  | 6.90 | negative  | Not Significant
#> 
#> Marginal effects estimated for Petal.Length
#> Type of slope was dY/dX

model <- mgcv::gam(Sepal.Width ~ s(Petal.Length, by = Species), data = iris)
slopes <- estimate_slopes(model,
  trend = "Petal.Length",
  by = c("Petal.Length", "Species"), length = 20
)
summary(slopes)
#> Average Marginal Effects
#> 
#> Group      | Start |  End | Direction | Confidence     
#> -------------------------------------------------------
#> setosa     |  1.00 | 1.62 | positive  | Not Significant
#> versicolor |  3.17 | 5.04 | positive  | Significant    
#> virginica  |  4.73 | 5.66 | positive  | Significant    
#> virginica  |  5.97 | 6.90 | positive  | Not Significant
#> 
#> Marginal effects estimated for Petal.Length
#> Type of slope was dY/dX

^{Created on 2025-01-29 with reprex v2.1.1}

[strengejacke]

m <- lm(Sepal.Width ~ Petal.Length * Petal.Width, data = iris)

modelbased::estimate_slopes(m, "Petal.Length", by = "Petal.Width") |> summary()
#> Average Marginal Effects
#> 
#> Start |  End | Direction | Confidence     
#> ------------------------------------------
#> 0.10  | 0.90 | negative  | Significant    
#> 1.17  | 1.70 | negative  | Not Significant
#> 1.97  | 2.23 | positive  | Not Significant
#> 2.50  | 2.50 | positive  | Significant    
#> 
#> Marginal effects estimated for Petal.Length
#> Type of slope was dY/dX

modelbased::estimate_slopes(m, "Petal.Length", by = "Petal.Width", length = 150) |> summary()
#> Average Marginal Effects
#> 
#> Start |  End | Direction | Confidence     
#> ------------------------------------------
#> 0.10  | 1.15 | negative  | Significant    
#> 1.16  | 1.69 | negative  | Not Significant
#> 1.71  | 2.44 | positive  | Not Significant
#> 2.45  | 2.50 | positive  | Significant    
#> 
#> Marginal effects estimated for Petal.Length
#> Type of slope was dY/dX

^{Created on 2025-01-29 with reprex v2.1.1}

[DominiqueMakowski]

Ill be travelling over this weekend so I'll get to this next week!

[strengejacke]

Yeah, I don't think this PR is too urgent.