2025-09-20
The R-squared for this tired and overly simple model is
model1 <- lm(hp ~ mpg, data = mtcars) round(summary(model1)$r.squared, 3)
## [1] 0.602
As we can see, MPG is a great predictor of horsepower. But, mtcars has so many other features for us to take advantage of!
modelNew <- lm(hp ~ mpg + disp + cyl, data = mtcars)
The R-squared for our new model is: 0.712. By using more data, we’ve significantly improved the model’s predictive power.
| Model | R-squared | Adjusted R-squared |
|---|---|---|
| Old (MPG) | 0.602 | 0.589 |
| New (MPG, Disp, Cyl) | 0.712 | 0.681 |
As we can see, our new model’s R-squared value is much higher, which means it explains a much larger proportion of the variation in horsepower.
Our new model is awesome and you should check out the ahem “shiny” new web app we’ve built to show it off. Click here NOW!
Obviously this was all written in jest and I’m just having fun. And as we all know - using more predictors isn’t actually always better. If we can, we should “keep it simple!”. Cheers!
