CrowdCleaner: Data cleaning for multi-version data on the web via crowdsourcing

摘要: Multi-version data is often one of the most concerned information on Web since this type usually updated frequently. Even though there exist some integration systems that try to maintain latest update version, maintained multi-version includes inaccurate and invalid due or delay errors. In demo, we present CrowdCleaner, a smart cleaning system for Web, which utilizes crowdsourcing-based approaches detecting repairing errors cannot be solved by traditional techniques. particular, CrowdCleaner blends active passive crowdsourcing methods together rectifying data. We demonstrate following four facilities provided CrowdCleaner: (1) an error-monitor find out items (e.g., submission date, price real estate, etc.) are wrong versions according reports from crowds, belongs strategy; (2) task-manager allocate tasks human workers intelligently; (3) smart-decision-maker identify answer crowds correct with methods; (4) whom-to-ask-finder discover users (or workers) should credible their records.