audit-er-results

Analyze entity resolution match results for coverage, false positives, and false negatives.

3|1|Updated Feb 7, 2025

Search Tags:#quality control #data quality #audit #entity resolution #false negatives

Ready to run

npx skills add https://github.com/thegoodparty/gp-data-platform --skill audit-er-results

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: audit-er-results
Source: https://github.com/thegoodparty/gp-data-platform/tree/main/matcha/.claude/skills/audit-er-results
Command: npx skills add https://github.com/thegoodparty/gp-data-platform --skill audit-er-results

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill provides comprehensive analysis for entity resolution match results, ensuring quality in coverage, identifying false positives and negatives, and guiding improvements.

Core Features & Use Cases

Coverage Analysis: Evaluate the match rates and cluster sizes to assess overall match quality.
Low-Confidence Review: Examine ambiguous match pairs with close probability scores for potential misclassifications.
False Negative Review: Investigate non-matches for missed opportunities, considering provider-specific issues.
Regression Check: Confirm existing matches remain accurate after any changes.
Recommendations Compilation: Provide actionable advice on blocking rules, comparison thresholds, and post-prediction filters.

Quick Start

Perform an audit of the entity resolution match results for the 'candidacy_stage' entity.

Frequently Asked Questions about audit-er-results

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

FAQPage Schema

How do I identify false negatives in entity resolution match results?▼

To identify false negatives in entity resolution, you can review non-matches to find missed opportunities by analyzing provider-specific issues and undermatching patterns using Python scripts.

What is entity resolution quality control and why is it needed?▼

Entity resolution quality control is the process of assessing match coverage, false positives, and false negatives to ensure data integrity and address common undermatching or overmatching issues in your results.

Do I need Python to audit entity resolution matches for data quality?▼

Yes, you need a Python environment to execute the scripts required to review match results, assess low-confidence pairs, and generate recommendations for blocking rules and comparison thresholds.

What's the best way to check for false positives in low-confidence match pairs?▼

The best way to check for false positives is to examine ambiguous match pairs with close probability scores to detect potential misclassifications and evaluate your comparison thresholds.

Why does my entity resolution coverage drop after updating blocking rules?▼

Coverage drops when new blocking rules cause undermatching by excluding valid pairs, requiring a regression check to confirm existing matches remain accurate and adjusting thresholds to restore quality.