Robustness of TabMNAR Metrics Across Missing Data Mechanisms and Domains via Synthetic and Real-World Benchmarking

Abstract

Abstract: Incomplete data is a persistent challenge in real-world datasets, often governed by complex and unobservable missing mechanisms. Simulating missingness has become a standard approach for understanding its impact on learning and analysis. However, existing tools are fragmented, mechanism-limited, and typically focus only on numerical variables, overlooking the heterogeneous nature of real-world tabular data. We present MissMecha, an open-source Python toolkit for simulating, visualizing, and evaluating missing data under MCAR, MAR, and MNAR assumptions. MissMecha supports both numerical and cat

Research Question

How robust are TabMNAR's metrics when applied to tabular data with different types of missing data mechanisms (MCAR, MAR, MNAR) across domains such as finance, healthcare, and social sciences, as measured by benchmarking against synthetic and real-world datasets?

Verification Level

Descriptive public verification status only; aggregate claim counts are public, but individual claim records are not exposed here.

Truth-Engine Gate Verdict

This record has not completed Gate 2 of the verification pipeline (a type-checked Lean4 proof for mathematical claims, or a sealed-sandbox reproduction for empirical claims). It is a literature synthesis only. VERIFIED requires an attached reproducible artifact (Lean4 proof source, or repro script and results) before this status can be set; it is not derived from review score or claim count.

Quality Tier

Descriptive public triage only; this tier does not alter current publication or DOI behavior.

Quality Dimensions

Automated triage signals derived from public fields; not human peer review or independent validation.

Correction Record

Public corrections are additive records. Current status does not claim the synthesis is error-free.

Provenance