A Data Quality Metric (DQM): How to Estimate The Number of Undetected Errors in Data Sets

TL;DR

This paper introduces new species estimators to accurately quantify remaining errors in datasets after cleaning, especially when errors are hard to detect and traditional methods are unstable, aiding decisions on further data cleaning efforts.

Contribution

The paper proposes novel species estimators specifically designed to handle false positives and negatives in crowd-sourced data cleaning, improving error estimation stability and convergence.

Findings

Existing estimators are unstable for this problem

Proposed estimators quickly converge on real-world datasets

New methods provide more reliable error quantification

Abstract

Data cleaning, whether manual or algorithmic, is rarely perfect leaving a dataset with an unknown number of false positives and false negatives after cleaning. In many scenarios, quantifying the number of remaining errors is challenging because our data integrity rules themselves may be incomplete, or the available gold-standard datasets may be too small to extrapolate. As the use of inherently fallible crowds becomes more prevalent in data cleaning problems, it is important to have estimators to quantify the extent of such errors. We propose novel species estimators to estimate the number of distinct remaining errors in a dataset after it has been cleaned by a set of crowd workers -- essentially, quantifying the utility of hiring additional workers to clean the dataset. This problem requires new estimators that are robust to false positives and false negatives, and we empirically show on three real-world datasets that existing species estimators are unstable for this problem, while our proposed techniques quickly converge.