Estimation of the Generalized Symmetry Point, an Optimal Cutpoint in Continuous Diagnostic Tests

Estimation of the cutpoint defined by the Generalized Symmetry point in a binary classification setting based on a continuous diagnostic test or marker. Two methods have been implemented to construct confidence intervals for this optimal cutpoint, one based on the Generalized Pivotal Quantity and the other based on Empirical Likelihood. Numerical and graphical outputs for these two methods are easily obtained.


CHANGES in `GsymPoint' VERSION 1.1.1
o	The R Journal reference was updated (including the correct authors)

o	The Pharmaceutical Statistics reference was updated

o	A new option "auto" was included in the main function to automatically select the most appropriate method of the two available "GPQ" and "EL" methods

o	Shapiro-Wilk's test p-values were fixed as NULL when the sample size falls outside of the interval [3,5000]

o	The AUC result for the GPQ method is now given directly, not in a nested list

o	The AUC result for the EL method is now given directly, not in a nested list

o	The same column names ("Value", "ll" and "ul") as in the GPQ method were included in the EL method 

o	The results of both methods EL and GPQ are now given under the list "Global"

o	The possibility of using a group variable with categories different from 0 and 1 was included

o	The type of the status variable to be used within gsym.point function was changed in the way that now the end-user is not forced to use the codification 0 and 1 for the status variable

o	Some options of the plot function were changed in the way that the end-user can change the title and the labels of the axes given by default

o	Output of the plot function was changed: 

		- The graphical display of the results includes the ROC-coordinates of the point estimates obtained
		- For datasets with ties now the accurate ROC curve is shown
		- The call to readline ("press return for next page") was avoided because it is problematic for non-interactive usage

o	The unnecessary code in the calc.empirical.AUC function was deleted (the confidence level argument and the count.zeros and count.neg variables defined inside)

o	Some unit tests were included using the testthat library in order to check that the correct ROC curves and AUC values are computed

o	A new parameter "verbose" was included in the main function to turn on/off the option of printing information on the screen while the code is running

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


1.1.1 by Mónica López-Ratón, 4 years ago

Browse source code at

Authors: Mónica López-Ratón [aut, cre] , Carmen Cadarso-Suárez [aut] , Elisa M. Molanes-López [aut] , Emilio Letón [aut]

Documentation:   PDF Manual  

GPL license

Depends on truncnorm, Rsolnp, ROCR

Suggests testthat

Suggested by WeightedROC.

See at CRAN