Dashboard Structural Analysis cloak.business SD1 LINKABILITY Case Study
← Previous Next →
cloak.business SD1 LINKABILITY
Case Study 6 of 30

From t-closeness to differential privacy and vice versa in data anonymization

J. Domingo-Ferrer, J. Soria-Comas (2015-12-16)

Research Source

From t-closeness to differential privacy and vice versa in data anonymization
J. Domingo-Ferrer, J. Soria-Comas · 2015-12-16 · Source: arxiv

k-Anonymity and ε-differential privacy are two mainstream privacy models, the former introduced to anonymize data sets and the latter to limit the knowledge gain that results from including one individual in the data set. Whereas basic k-anonymity only protects against identity disclosure, t-closeness was presented as an extension of k-anonymity that also protects against attribute disclosure.

Executive Summary

This research paper examines a critical privacy challenge related to LINKABILITY — the ability to connect two pieces of information to the same person.

cloak.business addresses this through 390+ entity types with 317 custom regex recognizers, processed in-memory on German servers with zero third-party data sharing.

Root Cause: SD1 — LINKABILITY

The ability to connect two pieces of information to the same person. This is the foundational operation that makes PII dangerous. Nearly every pain point is an expression of linkability being created, exploited, or failing to be broken.

Irreducible truth: You cannot have useful data that is completely unlinkable AND completely useful. The very features that make data informative make it linkable. This is not a bug — it is information theory. The information content of a dataset and its linkability are the same property measured differently.

The Solution: How cloak.business Addresses This

Detection Capabilities

cloak.business identifies 390+ entity types including text content, writing patterns, timestamps, posting metadata, timezone indicators. The dual-layer (317 custom regex + NLP) architecture uses 317 custom regex recognizers with context word analysis and confidence scoring 0.0–1.0 for structured identifiers and spaCy (25 languages) + Stanza (7 languages) + XLM-RoBERTa (16 languages) — all self-hosted for contextual references.

Anonymization Methods

Replace is recommended for this pain point: replacing original text content with anonymized alternatives disrupts the stylometric fingerprint that writing analysis algorithms depend on. Redact provides an alternative — removing text content entirely prevents any stylometric analysis though it reduces document utility. For scenarios requiring reversibility, Encrypt (AES-256-GCM) enables authorized recovery of original values.

Architecture & Deployment

The Desktop App (Windows 10+, Tauri/Rust) processes documents locally. Combined with zero-storage server architecture, PII is processed and immediately discarded.

Compliance Mapping

This pain point intersects with GDPR Article 4(1) personal data extends to indirectly identifying information including writing style.

cloak.business’s GDPR (Article 25 Privacy by Design), ISO 27001:2022 compliance coverage, combined with Germany only, no third-party transfers, ISO 27001:2022 certified hosting, provides documented technical measures organizations can reference in their compliance documentation and regulatory submissions.

Product Specifications

SpecificationValue
Platform VersionAnalyzer 6.9.1, Image Redactor 5.3.0
Entity Types390+ (519 documented)
Detection Layers317 custom regex + 3 NLP engines (all self-hosted)
Languages48 UI languages, 37 OCR language packs
Anonymization MethodsReplace, Redact, Mask, Hash (SHA-256), Encrypt (AES-256-GCM)
ArchitectureZero-storage microservices (in-memory only)
Integration PointsWeb App, Desktop, Office Add-in, MCP Server (9 tools), REST API
HostingGermany only, ISO 27001:2022, no third-party transfers
ComplianceGDPR Article 25, ISO 27001:2022
← Previous Next →