Skip to main content

Table 1 Feature importance in machine learning models for predicting perceptual sensitivity (d’) and metacognitive efficiency (log M-ratio) based on superiority rating scores

From: Overlapping yet dissociable contributions of superiority illusion features to Ponzo illusion strength and metacognitive performance

Item

Feature importance

d’

 #29

vain

0.203

 #21

dominating

0.128

 #05

moody

0.107

 #07

unimaginative

-0.119

 #08

warm

-0.194

 #13

insignificant

-0.237

 #35

reliable

-0.266

log M-ratio

 #19

irritable

0.299

 #48

irresponsible

-0.062

 #33

superficial

-0.101

 #32

practical

-0.116

 #50

helpful

-0.159

 #45

sociable

-0.323

  1. Note that response and predictor variables included in the machine learning models were standardized for each variable, so the model feature importance should be interpreted accordingly. For prediction performance, see Fig. 4