Table 7.
Classification results for binary classification models {G0+G1} versus G2+
| Dataset and classifier | AUCa | Accuracy | Recall | Precision | F1 | |
| Wearable |
|
|
|
|
|
|
| LRb | 0.599 (+/-0.306) | 0.730 (+/-0.185) | 0.599 (+/-0.252) | 0.518 (+/-0.294) | 0.551 (+/-0.275) | |
|
|
DTc | 0.768 (+/-0.209) | 0.730 (+/-0.184) | 0.750 (+/-0.210) | 0.682 (+/-0.262) | 0.678 (+/-0.238) |
|
|
RFd | 0.795 (+/-0.247) | 0.784 (+/-0.276) | 0.795 (+/-0.247) | 0.732 (+/-0.325) | 0.742 (+/-0.307) |
| RAI-HCe |
|
|
|
|
|
|
| LR | 0.842 (+/-0.306) | 0.649 (+/-0.156) | 0.610 (+/-0.174) | 0.514 (+/-0.264) | 0.524 (+/-0.214) | |
|
|
DT | 0.836 (+/-0.149) | 0.811 (+/-0.187) | 0.836 (+/-0.149) | 0.836 (+/-0.217) | 0.799 (+/-0.206) |
|
|
RF | 0.836 (+/-0.206) | 0.838 (+/-0.192) | 0.858 (+/-0.175) | 0.869 (+/-0.165) | 0.836 (+/-0.195) |
| Wearable + RAI-HC |
|
|
|
|
|
|
| LR | 0.838 (+/-0.234) | 0.703 (+/-0.172) | 0.676 (+/-0.200) | 0.657 (+/-0.281) | 0.626 (+/-0.231) | |
|
|
DT | 0.858 (+/-0.160) | 0.838 (+/-0.200) | 0.869 (+/-0.154) | 0.851 (+/-0.226) | 0.829 (+/-0.218) |
|
|
RF | 0.894 (+/-0.155) | 0.892 (+/-0.160) | 0.908 (+/-0.135) | 0.928 (+/-0.106) | 0.888 (+/-0.166) |
aAUC: area under the receiver operating characteristic curve.
bLR: logistic regression.
cDT: decision tree.
dRF: random forest.
eRAI-HC: Resident Assessment Instrument-Home Care.