[Verse 1]
When models multiply like rabbits in spring
Each claiming they're the perfect thing
But numbers never lie, they whisper truth
Through metrics that can cut right through
AIC penalizes every extra term
While BIC makes complexity squirm
Adjusted R-squared won't be fooled by bloat
These guardians decide which model gets your vote
[Chorus]
Lower AIC, that's the key
BIC agrees, sets you free
Adjusted R-squared climbing high
Pick the model that won't lie
Three wise judges, never wrong
Sing their statistical song
Model selection, crystal clear
Trust the metrics, persevere
[Verse 2]
Akaike Information Criterion stands guard
Against the overfitting yard
It counts parameters like coins you spend
Subtracts them from the likelihood blend
Bayesian Information goes one step more
Heavier penalties at its core
Sample size amplifies the punishment deal
For models that aren't quite real
[Chorus]
Lower AIC, that's the key
BIC agrees, sets you free
Adjusted R-squared climbing high
Pick the model that won't lie
Three wise judges, never wrong
Sing their statistical song
Model selection, crystal clear
Trust the metrics, persevere
[Bridge]
Regular R-squared fools you with its climb
Adding variables every single time
But adjusted version tells the honest tale
Shrinks when worthless predictors fail
Compare them side by side in neat arrays
Let mathematics guide your ways
[Verse 3]
Calculate each metric for every candidate
Line them up and don't be late
Smallest AIC wins the crown
BIC concurs, writes it down
Highest adjusted R-squared takes the prize
Three perspectives, eagle eyes
When they agree, your choice is gold
The story's beautifully told
[Chorus]
Lower AIC, that's the key
BIC agrees, sets you free
Adjusted R-squared climbing high
Pick the model that won't lie
Three wise judges, never wrong
Sing their statistical song
Model selection, crystal clear
Trust the metrics, persevere
[Outro]
Model selection mastery
Through the lens of these trusty three