Skip to main content
. 2021 Feb 9;4(3):ooaa069. doi: 10.1093/jamiaopen/ooaa069

Table 3.

Performance of models inferring SDH labels using 5-fold cross-validation

Algorithm Label Precision Recall F1 Support
SVM General financial insecurity 0.51 0.91 0.66 69
Employment/income insecurity 0.66 0.81 0.73 98
Housing insecurity 0.62 0.89 0.73 83
Poor social support 0.7 0.97 0.82 89
Insurance insecurity 0.62 0.84 0.71 73
Food insecurity 0.92 0.82 0.87 44
Xgboost General financial insecurity 0.84 0.71 0.77 69
Employment/income insecurity 0.85 0.76 0.8 98
Housing insecurity 0.88 0.69 0.77 83
Poor social support 0.85 0.93 0.89 89
Insurance insecurity 0.81 0.75 0.78 73
Food insecurity 0.93 0.86 0.89 44
KNN General financial insecurity 0.66 0.33 0.44 69
Employment/income insecurity 0.76 0.3 0.43 98
Housing insecurity 0.53 0.12 0.2 83
Poor social support 0.68 0.46 0.55 89
Insurance insecurity 0.56 0.3 0.39 73
Food insecurity 1 0.59 0.74 44
RF General financial insecurity 0.78 0.42 0.55 69
Employment/income insecurity 0.78 0.29 0.42 98
Housing insecurity 0.83 0.23 0.36 83
Poor social support 0.8 0.57 0.67 89
Insurance insecurity 0.7 0.1 0.17 73
Food insecurity 1 0.34 0.51 44