Thanks for the suggestion! Will try to do that in a next video. For now: you can save the predicted probabilities when you build your own model (for example logistic regression) by using the "save" button within the logistic regression menu, and than click "save predicted probabilities". It will than be saved as a new variable within your dataset. Good luck!
Excellent presentation Dr de Glas. I was hoping you can guide me regarding the method to calculate the concordance index concept (Harrell's c index) and the online prediction model development.
Thank you for the excellent presentation, I have som questions thoug: I have performed AUS-curves for 2 prediction models and want to present them nicely with CIs and p-value in the graf, can you show how to incorporate that in the graf? Also for showing differences between the models how is that done? and what is the name of that test?
Hi. Very useful video thanks. If the model have different risk thresholds for example i’m externally validate a model that have three risk thresholds(less than 15% low risk group), 15 -60% moderate and more than 60% high propability. Can i compute that in the roc curve?? Thank you
thankyou it helps me so much
The videos are so nice, focused and practical. Really thanks.
Thanks a lot for your kind words!
Very nice presentation as usual. Thank you so much for your great videos.
Thanks!!
Great explanation. Thank you!
Thanks for your excellent presentation.
It was a nice presentation. I wish you could also demonstrate the calculation of probabilities of each individual patient on SPSS. Thank you.
Thanks for the suggestion! Will try to do that in a next video. For now: you can save the predicted probabilities when you build your own model (for example logistic regression) by using the "save" button within the logistic regression menu, and than click "save predicted probabilities". It will than be saved as a new variable within your dataset. Good luck!
Thank you so much for sharing your valuable work
You are very welcome :)!
Excellent presentation Dr de Glas. I was hoping you can guide me regarding the method to calculate the concordance index concept (Harrell's c index) and the online prediction model development.
Can bmi consider as state variables n waist circumference or bp or whr as test variables? N how to arrive bmi cut off
Thank you for the excellent presentation, I have som questions thoug: I have performed AUS-curves for 2 prediction models and want to present them nicely with CIs and p-value in the graf, can you show how to incorporate that in the graf? Also for showing differences between the models how is that done? and what is the name of that test?
How abouit area under the curve for line graphs?
How to caluculate area under curve? please help.
Hi. Very useful video thanks. If the model have different risk thresholds for example i’m externally validate a model that have three risk thresholds(less than 15% low risk group), 15 -60% moderate and more than 60% high propability. Can i compute that in the roc curve??
Thank you
Great presentation, thank you! One question though, what if one filters out the lost to followup in SPSS, will the ROC curve work?
Wherevis the validation tutorial?