Loading...
Please wait, while we are loading the content...
Similar Documents
Statistical Comparison of Classifiers Using Area Under the ROC Curve
| Content Provider | Semantic Scholar |
|---|---|
| Author | Aslan, Özlem Yildiz, Olcay Taner Alpaydin, Ethem |
| Copyright Year | 2009 |
| Abstract | Statistical tests in the literature mainly use error rate for comparison. Receiver Operating Characteristics (ROC) curves and/or Area Under the ROC Curve (AUC) can also be used for comparing classifier performances under a spectrum of loss values. A ROC curve and hence an AUC value is calculated from one training/test pair and to average over randomness in folds, we propose to use k-fold cross-validation to generate a set of ROC curves and AUC values to which we can fit a distribution and test hypotheses on. Experiment results on 15 datasets using 5 different classification algorithms show that our proposed test using AUC values is to be preferred over the usual paired t test on misclassification errors because it can detect equivalences and differences which the error test cannot. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://haydut.isikun.edu.tr/cv/teknikrapor/2009-03.pdf |
| Alternate Webpage(s) | https://www.cmpe.boun.edu.tr/~aslan/technical-report.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
