h******e 发帖数: 1791 | 1 就是那个Receiver operating characteristic curve。谢谢。 |
f****r 发帖数: 1140 | 2 是看Logistic model fit的好不好? |
j*****e 发帖数: 182 | 3 One of its application is to compare different diagnostic methods. |
c*****l 发帖数: 631 | 4 就是看曲线下面积的大小,面积越大model越好,用来看model的好坏,也可以用来和
其他的model比较。 |
c*****r 发帖数: 2608 | 5 in diagnostic test
【在 h******e 的大作中提到】 : 就是那个Receiver operating characteristic curve。谢谢。
|
i******r 发帖数: 323 | 6 In logistic regression, ROC curve measure how good the model fits.
In SAS output, the value c is the area under the ROC curve. If 1, the
perfect model; if 0.5, then you should treat the data as random Bernoulli
trials, I would assume the goodness of fit test fails. |
i******r 发帖数: 323 | 7 In logistic regression, ROC curve measure how good the model fits.
In SAS output, the value c is the area under the ROC curve. If 1, the
perfect model; if 0.5, then you should treat the data as random Bernoulli
trials, I would assume the goodness of fit test fails. |