> Sys.time()
[1] "2014-10-21 10:57:40 EEST"

> summary(m1, digits = 2, style = "pmax")
Call: earth(formula=fmla, data=Dm)

ELTYY =
  6.6
  -    0.17 * pmax(0,       57 -  IKATUTK) 
  -    0.26 * pmax(0,  IKATUTK -       77) 
  -    0.21 * pmax(0,     JOB2 -    -0.22) 
  -    0.27 * pmax(0,    -0.71 -     JOB3) 
  +    0.18 * pmax(0,     ACT4 -     -1.3) 
  +    0.35 * pmax(0,     ACT5 -     0.41) 
  -    0.12 * pmax(0,     0.41 -     ACT5) 
  -   0.061 * pmax(0,        8 -  SOSVERK) 
  -  0.0012 * pmax(0,    SAIPV -       40) 
  +    0.02 * pmax(0,       40 -    SAIPV) 
  +    0.16 * pmax(0,   LAHTO1 -     -0.3) 
  -    0.33 * pmax(0,     -0.3 -   LAHTO1) 
  +    0.19 * pmax(0,      3.6 -   LAHTO2) 
  + 0.00012 * pmax(0, ELAKETUb -     1680) 
  - 0.00062 * pmax(0,     1680 - ELAKETUb) 

Selected 16 of 30 terms, and 10 of 20 predictors 
Importance: SAIPV, IKATUTK, SOSVERK, LAHTO1, ACT4, LAHTO2, JOB2, ELAKETUb, ...
Number of terms at each degree of interaction: 1 15 (additive model)
GCV 1.5    RSS 3516    GRSq 0.2    RSq 0.21

> evimp(m1, trim=FALSE);
              nsubsets   gcv    rss
SAIPV               15 100.0  100.0
IKATUTK             14  75.4   77.5
SOSVERK             13  64.5   67.5
LAHTO1              12  54.4   58.3
ACT4                11  47.6   52.0
LAHTO2               9  37.0   42.2
JOB2                 9  36.6   41.8
ELAKETUb             9  36.6   41.8
ACT5                 7  24.4   30.9
JOB3                 5  13.1   21.3
MIES-unused          0   0.0    0.0
EDUCI-unused         0   0.0    0.0
TUTKP-unused         0   0.0    0.0
DURURA-unused        0   0.0    0.0
JOB1-unused          0   0.0    0.0
JOB4-unused          0   0.0    0.0
ACT1-unused          0   0.0    0.0
ACT2-unused          0   0.0    0.0
ACT3-unused          0   0.0    0.0
ACT6-unused          0   0.0    0.0

> sink()