Roc Analysis Tools

Configuration File: roc_analysis_tools.json Tool Type: Local Tools Count: 1

This page contains all tools defined in the roc_analysis_tools.json configuration file.

Available Tools

ROC_analysis (Type: ROCAnalysisTool)

Deterministic ROC / AUC diagnostic-accuracy analysis for any binary classifier or continuous biom…

ROC_analysis tool specification

Tool Information:

  • Name: ROC_analysis

  • Type: ROCAnalysisTool

  • Description: Deterministic ROC / AUC diagnostic-accuracy analysis for any binary classifier or continuous biomarker. Give scores + 0/1 labels (inline arrays) or a CSV with score/label columns; returns AUC with a bootstrap 95% CI, the Youden-optimal cutoff and its sensitivity/specificity, optional metrics at a fixed cutoff, and a downsampled ROC curve. Pure NumPy compute (no scikit-learn, runs on a default install); no API key.

Parameters:

  • scores ([‘array’, ‘null’]) (optional) Continuous prediction scores (inline). Pair with ‘labels’.

  • labels ([‘array’, ‘null’]) (optional) Binary class labels aligned to scores (1=positive/0=negative, or any two values + ‘positive_label’).

  • csv_path ([‘string’, ‘null’]) (optional) Alternative to inline arrays: path to a CSV with score and label columns

  • score_col ([‘string’, ‘null’]) (optional) Score column name in csv_path (default ‘score’)

  • label_col ([‘string’, ‘null’]) (optional) Label column name in csv_path (default ‘label’)

  • positive_label ([‘integer’, ‘string’, ‘null’]) (optional) Which label value is the positive class (default: 1, or the larger of two classes)

  • cutoff ([‘number’, ‘null’]) (optional) Optional fixed cutoff; also reports sensitivity/specificity at score >= cutoff

Example Usage:

query = {
    "name": "ROC_analysis",
    "arguments": {
    }
}
result = tu.run(query)