concept-drift-analysis

Correlate sliced AUC/F1 metrics with ground-truth labels to diagnose concept drift.

5|Updated Nov 18, 2025

Search Tags:#mlops #drift #rca #slicing #ground-truth #performance-analysis #concept-drift

Ready to run

npx skills add https://github.com/DuqueOM/ML-MLOps-Portfolio --skill concept-drift-analysis

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: concept-drift-analysis
Source: https://github.com/DuqueOM/ML-MLOps-Portfolio/tree/main/.devin/skills/concept-drift-analysis
Command: npx skills add https://github.com/DuqueOM/ML-MLOps-Portfolio --skill concept-drift-analysis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Detect and diagnose concept drift in performance alerts by linking sliced performance metrics with ground-truth labels, enabling targeted remediation.

Core Features & Use Cases

Read latest performance reports and slice-level metrics.
Correlate AUC/F1 drops with ground-truth signals to distinguish data drift from model degradation.
Provide actionable next steps for retraining or data-quality fixes, with traceable decision records.

Quick Start

Run concept-drift-analysis on the latest performance report to identify the root cause of a sliced AUC drop.

Frequently Asked Questions about concept-drift-analysis

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I diagnose concept drift from performance alerts using sliced metrics?▼

Diagnose concept drift by correlating per-slice AUC and F1 drops with ground-truth labels to distinguish data drift from model degradation. This analysis cross-references performance reports with drift reports to pinpoint the root cause.

What is the difference between data drift and model degradation when analyzing root cause?▼

Data drift indicates shifting input distributions, whereas model degradation signifies weakened predictive power. Cross-referencing sliced performance metrics with ground-truth signals helps distinguish between the two to determine the correct root cause.

How do I investigate a sliced AUC drop using performance reports and ground-truth labels?▼

Investigate a sliced AUC drop by reading performance.json and baseline_metrics.json, then correlating those metrics with ground-truth labels. Cross-referencing an optional drift_report.json provides a clear root cause analysis and actionable next steps.

Do I need a separate drift report to perform root-cause analysis on concept drift?▼

A separate drift report is optional for root-cause analysis. You can diagnose concept drift using performance.json and baseline_metrics.json, but cross-referencing an optional drift_report.json improves the accuracy of distinguishing data drift from model degradation.

What next steps should I take after identifying concept drift in my sliced metrics?▼

After identifying concept drift, recommended next steps involve targeted remediation actions such as model retraining or implementing data-quality fixes. This process yields traceable decision records for your MLOps pipeline.

Can I use this concept drift analysis without baseline metrics?▼

Baseline metrics from baseline_metrics.json are required to evaluate concept drift. Comparing current sliced AUC and F1 performance against these baseline metrics is essential to detect drops and determine the root cause.