Page 407 - Applied Statistics with R
P. 407
16.2. SELECTION PROCEDURES 407
AIC vs Model Complexity
282
280
AIC
278
276
274
2 3 4 5 6 7 8 9
p, number of parameters
We could easily repeat this process for BIC.
RSS
BIC = log ( ) + log( ) .
hipcenter_mod_bic = n * log(all_hipcenter_mod$rss / n) + log(n) * (2:p)
which.min(hipcenter_mod_bic)
## [1] 1
all_hipcenter_mod$which[1,]
## (Intercept) Age Weight HtShoes Ht Seated
## TRUE FALSE FALSE FALSE TRUE FALSE
## Arm Thigh Leg
## FALSE FALSE FALSE
hipcenter_mod_best_bic = lm(hipcenter ~ Ht, data = seatpos)

