Predicting four outcomes

N Foldnes

Variable importance

The following regression models were trained

[1] "lm"     "glmnet" "knn"    "pls"    "gbm"

and their combined variable importance was calculated for each of four outcomes The variable importance, as ranked by combining these five methods:

First Plot of variable importance

Second Plot of variable importance.

Multiple regression for PAP

term estimate std.error statistic p.value
(Intercept) 0.00 0.08 0.00 1.00
Samtykke_L -0.20 0.09 -2.07 0.04
hap_L 0.20 0.08 2.40 0.02
Ref_L 0.11 0.09 1.29 0.20
Sinne_F -0.30 0.09 -3.36 0.00
Adverb_F 0.19 0.09 2.02 0.05
Konj_F 0.15 0.09 1.63 0.11
Subst_F 0.20 0.09 2.11 0.04
Verb_F -0.51 0.16 -3.27 0.00
Funksjon_F 0.32 0.12 2.69 0.01
Kausalt_F 0.12 0.09 1.33 0.19
WC_F -0.23 0.10 -2.21 0.03
De_F 0.11 0.09 1.29 0.20
Hverb_F 0.25 0.14 1.73 0.09
Se_F 0.11 0.09 1.12 0.27
Angst_F 0.16 0.09 1.85 0.07
Prep_F -0.18 0.10 -1.82 0.07

Multiple regression for PAP: Plot

Multiple regression for ENG

term estimate std.error statistic p.value
(Intercept) 0.00 0.08 0.00 1.00
hap_L 0.19 0.08 2.36 0.02
Stotring_L -0.20 0.08 -2.49 0.01
Inklusjon_L -0.10 0.08 -1.28 0.20
Negemo_L -0.21 0.08 -2.58 0.01
Hanhun_L 0.11 0.08 1.40 0.16
Biologisk_L 0.22 0.08 2.73 0.01
Sikker_L 0.09 0.08 1.15 0.25
Rom_L 0.10 0.08 1.14 0.26
Pronomen_F 0.34 0.13 2.70 0.01
Hverb_F 0.19 0.10 1.86 0.07
Naatid_F 0.18 0.10 1.82 0.07
Bevegelse_F 0.14 0.08 1.80 0.08
Funksjon_F -0.20 0.13 -1.49 0.14
sad_F -0.27 0.08 -3.56 0.00
Angst_F 0.18 0.08 2.24 0.03

Multiple regression for ENG: Plot

Multiple regression for MAX

term estimate std.error statistic p.value
(Intercept) 0.00 0.08 0.00 1.00
Samtykke_L -0.20 0.09 -2.21 0.03
Eksklusjon_L 0.17 0.09 2.00 0.05
Ref_L 0.15 0.09 1.80 0.08
Tid_L 0.13 0.09 1.57 0.12
Angst_L -0.17 0.09 -1.95 0.05
Sinne_F -0.21 0.08 -2.48 0.02
Subst_F 0.16 0.09 1.85 0.07
Konj_F 0.13 0.09 1.55 0.12
De_F 0.14 0.09 1.52 0.13
Funksjon_F 0.13 0.09 1.48 0.14
Persepsjon_F -0.14 0.09 -1.57 0.12
Adverb_F 0.22 0.09 2.53 0.01

Multiple regression for MAX: Plot

Multiple regression for SAT

term estimate std.error statistic p.value
(Intercept) 0.00 0.07 0.00 1.00
Fortid_L 0.10 0.08 1.25 0.22
Sixltr_L -0.17 0.08 -2.09 0.04
Kausalt_L -0.21 0.08 -2.79 0.01
WC_L 0.21 0.08 2.52 0.01
Hoere_L 0.16 0.08 2.00 0.05
hap_L 0.25 0.08 3.23 0.00
Biologisk_L 0.20 0.08 2.58 0.01
Sinne_L 0.09 0.08 1.12 0.26
Subst_L 0.11 0.09 1.19 0.24
Bevegelse_F 0.26 0.08 3.39 0.00
Narrative_F 0.23 0.09 2.57 0.01
Fortid_F 0.22 0.08 2.79 0.01
Stotring_F -0.09 0.08 -1.13 0.26
Kvantitet_F -0.12 0.08 -1.50 0.14
sad_F -0.17 0.08 -2.15 0.03
Se_F 0.16 0.08 2.07 0.04
Sixltr_F 0.11 0.10 1.11 0.27

Multiple regression for SAT: Plot

ALL COMBINED