Table 4 Comparison of LR, LASSO, and RF models for predicting CASP-12 difference between Wave 5 and Wave 4 using data from Wave 4 as a baseline. Mean (sd) performance was computed on 10 different training and test set splits
From: A model to forecast the two-year variation of subjective wellbeing in the elderly population
Metric | Dataset | LR | LASSO | RF |
|---|---|---|---|---|
accuracy | training set | 65% (0.142) | 65% (0.199) | 65% (0.003) |
test set | 64% (0.581) | 64% (0.56) | 64% (0.799) | |
AUC | training set | 0.651 (0.001) | 0.650 (0.002) | 1 (0) |
test set | 0.640 (0.006) | 0.639 (0.006) | 0.636 (0.01) | |
F1 score | training set | 0.682 (0.002) | 0.681 (0.002) | 1 (0) |
test set | 0.668 (0.006) | 0.667 (0.006) | 0.706 (0.007) |