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

Brain Informatics

Table 4 Feature significance with respect to PHQ9 using the permutation importance approach

PHQ9	Regression	KNN	SVM	Decision tree	Random forest	Gradient boost	Average
Gender	0.040	0.063	0.045	0.058	0.054	0.054	0.052
Ethnics	0.036	0.061	0.025	0.056	0.044	0.031	0.042
Education	0.039	0.069	0.026	0.059	0.052	0.039	0.047
Relationship	0.032	0.065	0.040	0.058	0.047	0.035	0.046
Adversity	0.080	0.056	0.059	0.054	0.067	0.068	0.064
Exercise	0.148	0.080	0.165	0.085	0.111	0.133	0.120
Alcohol	0.009	0.044	0.019	0.035	0.029	0.023	0.026
Tobacco	0.053	0.041	0.022	0.046	0.038	0.047	0.041
Relation-Impact	0.191	0.079	0.210	0.096	0.133	0.176	0.148
Communication	0.113	0.085	0.097	0.077	0.092	0.123	0.098
Therapy	0.095	0.062	0.092	0.065	0.055	0.067	0.073
Medication	0.053	0.055	0.064	0.069	0.056	0.049	0.058
Health-service	0.046	0.065	0.053	0.062	0.059	0.044	0.055
Social-distancing	0.017	0.028	0.013	0.031	0.028	0.042	0.026
Risk-group	0.010	0.041	0.015	0.045	0.040	0.017	0.028
Living-group	0.029	0.063	0.038	0.064	0.058	0.033	0.047
Contract-risk	0.008	0.044	0.017	0.042	0.037	0.020	0.028