A summary of the output of the logsplinefit quantlet:
Contents of summary
[ 1,] "=============================================================="
[ 2,] " LogSpline summary "
[ 3,] "--------------------------------------------------------------"
[ 4,] " knots loglik AIC penalty "
[ 5,] " min max "
[ 6,] " 4 -422.73 873.469 309.29 +INF "
[ 7,] " 5 -268.09 571.180 15.74 309.29 "
[ 8,] " 6 -260.22 562.437 2.09 15.74 "
[ 9,] " 7 -260.14 569.282 +NAN +NAN "
[10,] " 8 -258.12 572.249 0.36 2.09 "
[11,] " 9 -257.95 578.891 0.31 0.36 "
[12,] " 10 -257.79 585.577 0.15 0.31 "
[13,] " 11 -257.71 592.430 0.00 0.15 "
[14,] " 12 -257.71 599.425 0.00 0.00 "
[15,] " 13 -257.71 606.425 0.00 0.00 "
[16,] "--------------------------------------------------------------"
[17,] " the present optimal number of knots is 6 "
[18,] "--------------------------------------------------------------"
[19,] " penalty(AIC) was 7 "
[20,] " the default (BIC) would have been 5.606 "
[21,] "--------------------------------------------------------------"
[22,] " models with fewer than 4 knots can be fitted, but they are "
[23,] " not optimal for the present choice of penalty - choose "
[24,] " penalty in logsplinefit larger to see these fits "
[25,] "=============================================================="