A receiver operating characteristics (ROC) curve is a graphical approach which assess the performance of a binary classifier system. The ROC curve analysis is widely used in medicine, radiology, biometrics and various application of machine learning.

Here we developed an easy way to carry out ROC analysis. This application creates ROC curves, calculates area under the curve (AUC) values and confidence intervals for the AUC values, and performs multiple comparisons for ROC curves in a user-friendly, up-to-date and comprehensive way. Moreover, easyROC computes and compares partial AUCs. It can also perform sample size calculation.

An important feature of this application is to determine cut-off values especially for diagnostic tests. For this task, we made use of OptimalCutpoints package (Lopez-Raton et al, 2014) of R [1].

If you use easyROC web-tool in your researches, please cite easyROC as **Goksuluk D, Korkmaz S, Zararsiz G, Karaağaoğlu AE (2016). easyROC: An Interactive Web-tool for ROC Curve Analysis Using R Language Environment. The R Journal, 8(2):213-230. **Click here for the paper.

**Disclaimer:** This server is intended for research purposes only, not for clinical or commercial use.
It is a non-profit service to the scientific community, provided on an "AS-IS " basis without any warranty,
expressed or implied. The authors can not be held liable in any way for the service provided here.

Hacettepe University Faculty of Medicine Department of Biostatistics

dincer.goksuluk@hacettepe.edu.tr

Hacettepe University Faculty of Medicine Department of Biostatistics

selcuk.korkmaz@hacettepe.edu.tr

Hacettepe University Faculty of Medicine Department of Biostatistics

gokmen.zararsiz@hacettepe.edu.tr

** Version 1.3.1 (July 25, 2016)**

(1) Minor fixes: Added feature to keep only pairwise complete data. Missing cases are now removed before ROC curve analysis which causes to null return in ROC statistics.

** Version 1.3 (July 25, 2016)**

(1) Support for prametric ROC curve approximation.

(2) Minor bug fixes and improvements.

(3) Minor changes in user interface.

** Version 1.2 (May 6, 2016)**

(1) User manual added.

(2) Minor bug fixes and improvements.

(3) Re-checked the package dependencies.

** Version 1.1 (June 23, 2015) **

(1) Partial AUC feature has been added.

(2) Sample size calculation tab has been added.

(3) Minor improvements and bug fixes.

** Version 1.0 (March 19, 2015)**

(1) Initial version has been released.

**MLViS: a machine learning-based virtual screening tool**

**MVN: a web-tool for assessing multivariate normality **

**DDNAA: Decision support system for differential diagnosis of nontraumatic acute abdomen **

Load your data set in *.txt file format using this tab.

Rows must represent the observations and each column must represent the variables.

First row must be a header which indicates the variable names.

Use this tab to perform ROC curve analysis. easyROC supports both parametric and nonparametric approximations for ROC curve analysis.

First select marker(s), where all names of the variables, except the status variable, will be imported automatically by the tool.

Once the markers are selected, the direction should be defined. By default, higher values indicate higher risks.

Under

**Statistics**subtab, you can get area under the curve (AUC) value and its standard error, confidence interval and statistical significance, instantly.

One may select one of parametric or nonparametric approximations under

**Advanced options**checkbox (By default, the nonparametric approach is selected). The standart errors can be estimated using one of the proposed methods. Likewise, users can select a method for confidence inerval estimation. Moreover, one can also change the type I error (Default is 0.05).

Furthermore, the ROC curve plot can be obtained under this tab. There are plenty of options under the

**Plot options**checkbox, such as font type, axis label and colour etc.

Each false positive and true positive points can be found under **ROC Coordinates** subtab for each marker.

**Multiple Comparisons** subtab can be used to perform pairwise statistical comparisons for two or more ROC curves.

The comparison methods can be changed under

*Multiple Comparison Method*option. Available methods are Bonferroni (by default), False discovery rate and none (i.e no adjustment on multiple tests).

**Partial AUC** subtab gives partial AUC value(s) for specified ranges based on both sensitivity and specificity.

Users can determine optimal cut-off points for their marker(s) using this tab.

First, a ROC curve analysis has to be done in order to use this option.

Then, one of the markers, which are used for ROC curve analysis, can be selected to determine the optimal cut-off points.

One can select one of 34 methods for optimal cut-off point determination.

These methods can be found in the OptimalCutpoints package of R.

Several graphs, including ROC curve with the optimal cut-off point, Sensitivity & Specificity Curve and Distribution graphs, can be created as well.

Sample size calculation for ROC curve analysis can be implemented under this tab.

There are three different options for sample size calculation.

One can perform a sample size calculation for a single diagnostic test, comparison of two diagnostic tests or noninferiority of a new test to a standard test.

Please see Obuchowski, 2005 for further details about the methods.