The Receiver operating characteristic (ROC) curve, is a widely used statistical tool that can quantitatively evaluate the accuracy of Binary classification outcomes. The accuracy of the outcomes of different classifiers can also be compared statistically, to determine whether thte difference between the outcomes of two classification methods is statistically significant.

ROC analysis an important tool for the assessment of the accuracy of diagnostic tests, because diagnostic results are often either positive or negative judgments on the proposition that the test results of the subject are positive. The essence is a classification problem containing two possible opposite outcomes.

Although there is commercial statistical software providing ROC analysis functions, it is still not easy for scientific researchers, especially doctors engaged more in clinical work, to correctly use the ROC analysis functions and to correctly interpret the results of ROC analysis.

The University of Chicago made an important contribution to the methodology of ROC analysis and developed a set of ROC analysis algorithms for the assessment of diagnostic accuracy. Unfortunately, the University of Chicago's ROC software is now out of maintainance due to a lack of continued funding support. This website inherits the ROC software developed by Dr. Charles Metz and his collegues, and plans to develop the application programming interface for different programming languages gradually. It is also on the to-do list to display data using visualization techniques for better data understanding and to interpret the ROC results for more technical viewpoint.

Again, due to the lack of funding support, this site will not be updated very frequently, but the feedbacks and suggestions will be screened and solved one by one according to the priority.