You are here: CTSPedia>CTSpedia Web>StatisticalTools>SasTools>ROCcutoff (18 Mar 2010, RuiChen)EditAttach
-- RuiChen - 17 Mar 2010

MacroHeaderForm edit

Macro Name %ROCcutoff
Description The macro is used to find a cutoff for a continuous predictor, which maximizes the linear combination of specificity and sensitivity when the response variable is binary. A final logistic regression will fit a model using the resulting cutoff.
Macro Parameters DSName = name of SAS data set
Dir = directory the result file saved.
ID = name of unique id for each observation.
Outcome = name of outcome binary variable.
PredictNum = name of numeric predictor variable.
Format = format of the outcome variable.
WSpec = any number within the interval [0,1], which maximizes WSpec*specificity+(1-WSpec)*sensitivity
e.g. if WSpec=1, the cutpoint is found to maximize the specificity
if WSpec=0, the cutpoint is found to maximize the sensitivity
if WSpec=0.5, the cutpoint is found to maximize the sum of the specificity and sensitivity
OutForm = HTML or RTF, Format of output file;
Programmer Rui Chen
Revision March 18, 2010
Disclaimer This SAS macro is developed by Rui Chen in the Department of Biostatistics and Computational Biology at the University of Rochester under the auspice of the CTSI BREAD with the intent to facilitate data analyses and reporting in clinical and related research. It is copyrighted by Dr. Chen, March, 2010 and distributed for free public access.
Example(s) SampleRun.sas: example
ExampleRun.rtf: output
Special Features Provide a model ROC
Called Macros N/A
Notes

See Also

Edit | Attach | Print version | History: r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r2 - 18 Mar 2010 - 08:52:56 - RuiChen
 
Copyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding CTSPedia? Send feedback