How does the newer ROC Analysis procedure differ from the older ROC Curve procedure?
I'd like to have SPSS or somebody to highlight, point by point, what are "new and different" features in the ROC Analysis compared to the ROC Curve.
The documentation seem to miss the explanation.

------------------------------
Kirill Orlov
------------------------------

#SPSSStatistics

Original Message:
Sent: 10/13/2021 8:11:00 AM
From: Kirill Orlov
Subject: ROC Analysis vs ROC Curve

How does the newer ROC Analysis procedure differ from the older ROC Curve procedure?
I'd like to have SPSS or somebody to highlight, point by point, what are "new and different" features in the ROC Analysis compared to the ROC Curve.
The documentation seem to miss the explanation.

------------------------------
Kirill Orlov
------------------------------

#SPSSStatistics

Thank you, Rick. I can see it. However, what I would like to recommend SPSS documentation to do - is to add a paragraph commenting on the differences between the new and the old ROC procedures. To add the paragraph both in Command Synrax Reference and in the Related Procedures subsection in Help. And a line about it in Case Studies - as well.

------------------------------
Kirill Orlov
------------------------------

Original Message:
Sent: Wed October 13, 2021 09:07 AM
From: Rick Marcantonio
Subject: ROC Analysis vs ROC Curve

Hello, Kirill.

The old ROC:



The "new" ROC (ROC ANALYSIS), where I have highlighted the differences in syntax:



So, DESIGN is where most of the difference lies. For more details, see:

https://www.ibm.com/docs/en/spss-statistics/SaaS?topic=analysis-design-subcommand-roc-command, where it talks about paired-samples (PAIR, obviously) and the various options for SELECT.

There is also the Precision-Recall interpolation feature in the newer procedure (/PLOT).

Original Message:
Sent: 10/13/2021 8:11:00 AM
From: Kirill Orlov
Subject: ROC Analysis vs ROC Curve

How does the newer ROC Analysis procedure differ from the older ROC Curve procedure?
I'd like to have SPSS or somebody to highlight, point by point, what are "new and different" features in the ROC Analysis compared to the ROC Curve.
The documentation seem to miss the explanation.

------------------------------
Kirill Orlov
------------------------------

#SPSSStatistics

Original Message:
Sent: 10/13/2021 9:27:00 AM
From: Kirill Orlov
Subject: RE: ROC Analysis vs ROC Curve

Thank you, Rick. I can see it. However, what I would like to recommend SPSS documentation to do - is to add a paragraph commenting on the differences between the new and the old ROC procedures. To add the paragraph both in Command Synrax Reference and in the Related Procedures subsection in Help. And a line about it in Case Studies - as well.

------------------------------
Kirill Orlov
------------------------------

Original Message:
Sent: Wed October 13, 2021 09:07 AM
From: Rick Marcantonio
Subject: ROC Analysis vs ROC Curve

Hello, Kirill.

The old ROC:



The "new" ROC (ROC ANALYSIS), where I have highlighted the differences in syntax:



So, DESIGN is where most of the difference lies. For more details, see:

https://www.ibm.com/docs/en/spss-statistics/SaaS?topic=analysis-design-subcommand-roc-command, where it talks about paired-samples (PAIR, obviously) and the various options for SELECT.

There is also the Precision-Recall interpolation feature in the newer procedure (/PLOT).

Original Message:
Sent: 10/13/2021 8:11:00 AM
From: Kirill Orlov
Subject: ROC Analysis vs ROC Curve

How does the newer ROC Analysis procedure differ from the older ROC Curve procedure?
I'd like to have SPSS or somebody to highlight, point by point, what are "new and different" features in the ROC Analysis compared to the ROC Curve.
The documentation seem to miss the explanation.

------------------------------
Kirill Orlov
------------------------------

#SPSSStatistics