i want to knoe in details hoe to calcule the cut off value for roc curve and hoe to make this rock curve and the calculation formula for sensistivity and specificity
Hi,
SPSS software will helpful to measure sensitivity, specificity of your methodology or your input with compared data differentiation outcome with area under curve (AUC) or cut-off value.
The best cutt-off values are based on the calculation of Youden index which takes into account the sensitivity and the specificity. You take the greater Youden index.
Hi Noha,
The receiver operating characteristic (ROC) curve is a procedure that can produce both tabular and graphical output to aid in the assessment of a cutoff value used to create a dichotomous variable. Explain what the specific mean of the AUC. A Chi-square test provides a p-value associated with the null hypothesis that AUC equals 0.500. This test is also significant and it is recommended to mention its results. A large Youden index may be one criterion for deciding an appropriate cutoff score. Choose the very first row of the output, which has the highest Youden index statistic (provide the data and comment).
The following link will show you a practical example
Just to add to what other authors said, to determine the best cutoff value, you look at the one with the highest likelihood which will correspond to the best sensitivity and best specificity. If you are using graphPad, you can easily get this from the tables already calculated for you.
Most statistical tools such as the graphPad usually give the sensitivity and specificity automatically when the ROC curve is drawn and for each pair value for sensitivity and specificity a cutoff value. The problem is generally to choose the cutoff value with the best sensitivity and specificity since choosing a very high sensitivity will usually give low specificity. So what you can do is choose the cutoff with the best sensitivity and specificity pair.
- Badges
- Science topic