Using smoothing parameters to avoid overfitting
The smoothing parameter balances between likelihood and wiggliness to optimize model fit. Here, you'll examine smoothing parameters and will fit models with different fixed smoothing parameters.
This exercise is part of the course
Nonlinear Modeling with Generalized Additive Models (GAMs) in R
Exercise instructions
- View the value of the smoothing parameter (\(\lambda\)) of the provided
gam_mod
model by extracting thesp
value from the model. - Fit two models to the
mcycle
data withaccel
as a smooth function oftimes
and a smoothing parameter of:- 0.1
- 0.0001
- Visualize both models.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
library(mgcv)
# Extract the smoothing parameter
gam_mod <- gam(accel ~ s(times), data = mcycle, method = "REML")
___
# Fix the smoothing parameter at 0.1
gam_mod_s1 <- gam(accel ~ s(times), data = mcycle, sp = ___)
# Fix the smoothing parameter at 0.0001
gam_mod_s2 <- gam(___)
# Plot both models
par(mfrow = c(2, 1))
plot(gam_mod_s1, residuals = TRUE, pch = 1)
plot(gam_mod_s2, residuals = TRUE, pch = 1)