Research Article
Defining an Optimal Cut-Point Value in ROC Analysis: An Alternative Approach
Table 5
Relative bias and mean square error (MSE) of all methods. The normal homoscedastic unbalanced scenario
.
| | | Sample sizes | Minimum value | Youden index | Concordance probability | Point closest-to-(0-1) corner | Index of Union | | | Relative bias | MSE | Relative bias | MSE | Relative bias | MSE | Relative bias | MSE | Relative bias | MSE |
| 0.25 | 0.39 | 50 | 100 | 0.1750 | 0.5041 | 0.2046 | 0.1813 | 0.1042 | 0.0520 | 0.1007 | 0.0341 | 0.0686 | 0.0075 | 0.46 | 50 | 150 | 0.2986 | 0.5124 | 0.1683 | 0.1708 | 0.1142 | 0.0515 | 0.1324 | 0.0338 | 0.0923 | 0.0079 | 0.51 | 50 | 200 | 0.2955 | 0.5517 | 0.1420 | 0.1741 | 0.1483 | 0.0484 | 0.1544 | 0.0320 | 0.1059 | 0.0074 |
| 0.52 | 0.75 | 50 | 100 | 0.0361 | 0.1963 | 0.1061 | 0.0921 | 0.0572 | 0.0469 | 0.0481 | 0.0256 | 0.0535 | 0.0161 | 0.87 | 50 | 150 | 0.0613 | 0.1922 | 0.1141 | 0.0904 | 0.0588 | 0.0435 | 0.0746 | 0.0255 | 0.0597 | 0.0162 | 0.96 | 50 | 200 | 0.0601 | 0.2060 | 0.0926 | 0.0785 | 0.0690 | 0.0399 | 0.0778 | 0.0233 | 0.0696 | 0.0159 |
| 0.84 | 1.09 | 50 | 100 | 0.0006 | 0.0952 | 0.0411 | 0.0597 | 0.0157 | 0.0414 | 0.0219 | 0.0219 | 0.0162 | 0.0255 | 1.23 | 50 | 150 | 0.0030 | 0.0956 | 0.0498 | 0.0580 | 0.0313 | 0.0423 | 0.0375 | 0.0232 | 0.0328 | 0.0259 | 1.33 | 50 | 200 | 0.0028 | 0.0996 | 0.0440 | 0.0536 | 0.0345 | 0.0374 | 0.0409 | 0.0201 | 0.0342 | 0.0248 |
| 1.28 | 1.50 | 50 | 100 | | 0.0581 | 0.0350 | 0.0460 | 0.0177 | 0.0390 | 0.0225 | 0.0254 | 0.0195 | 0.0325 | 1.63 | 50 | 150 | | 0.0528 | 0.0378 | 0.0479 | 0.0254 | 0.0414 | 0.0298 | 0.0242 | 0.0270 | 0.0342 | 1.72 | 50 | 200 | 0.0007 | 0.0507 | 0.0469 | 0.0477 | 0.0380 | 0.0422 | 0.0386 | 0.0257 | 0.0368 | 0.0356 |
|
|
~, ~, and was taken as 0.51, 1.05, 1.68, and 2.56, respectively.
|