Skip to main content

Table 7 Feature significance with respect to EQ5D5L using the permutation importance approach

From: Investigating the mental health of university students during the COVID-19 pandemic in a UK university: a machine learning approach using feature permutation importance

EQ5D5L

Regression

KNN

SVM

Decision tree

Random forest

Gradient boost

Average

Gender

0.016

0.054

0.030

0.054

0.048

0.046

0.042

Ethnics

0.008

0.060

0.025

0.060

0.057

0.026

0.039

Education

0.013

0.062

0.027

0.051

0.046

0.025

0.037

Relationship

0.012

0.064

0.028

0.047

0.046

0.032

0.038

Adversity

0.045

0.049

0.075

0.068

0.047

0.048

0.055

Exercise

0.096

0.062

0.091

0.060

0.062

0.099

0.078

Alcohol

0.021

0.038

0.023

0.029

0.030

0.037

0.030

Tobacco

0.032

0.042

0.031

0.029

0.037

0.024

0.033

Relation-Impact

0.133

0.075

0.153

0.070

0.090

0.132

0.109

Communication

0.069

0.070

0.063

0.057

0.063

0.053

0.063

Therapy

0.180

0.091

0.150

0.106

0.115

0.144

0.131

Medication

0.117

0.067

0.082

0.091

0.077

0.070

0.084

Health-service

0.108

0.076

0.090

0.079

0.092

0.092

0.090

Social-distancing

0.008

0.026

0.012

0.030

0.023

0.033

0.022

Risk-group

0.123

0.061

0.070

0.066

0.084

0.091

0.083

Living-group

0.016

0.055

0.030

0.068

0.053

0.033

0.043

Contract-risk

0.004

0.046

0.019

0.032

0.029

0.015

0.024

  1. The top-5 most significant features highlighted in bold