ROC Curve

Evaluate how well a continuous score separates two classes (e.g., Torpor vs Active). The area under the curve (AUC) is computed and annotated.

Visualization Parameters

Upload your data and generate the roc curve

Upload Data File

TSV/CSV: Score (numeric) + Label (0/1 or class names)

Required Files

scores.tsv — two columns: numeric score + class label

Format

Row = one sample
Score: numeric predictor value (auto-detected)
Label: 0/1 or text classes — Positive/Negative, Torpor/Active, case/control are recognized
Higher score should indicate the positive class
AUC is computed by the trapezoid rule and printed on the plot

Example Data

Score	Label
10.576	Torpor
6.512	Active
9.045	Torpor
6.254	Active

ROC Curve

Upload data and click Generate, or run with example data

Documentation

Input format: two columns — a numeric score and a class label. Labels may be 0/1 or text (Positive/Negative, Torpor/Active, ...); both columns are detected automatically.

Output: ROC curve (sensitivity vs 1 − specificity) with the diagonal reference, shaded area, and the AUC value (trapezoid rule).

Typical use: assess candidate biomarkers or machine-learning scores that classify physiological states (e.g., torpor entry).