基于Web信息的关系型信息错误自动检测与修复技术研究综述

摘要

Information quality has become an important issue in many application areas.Automatically detecting and correcting information errors has proven to be an effective way to improve information quality in most information systems.Integrating information from the World Wide Web (WWW) can help us overcome the shortcomings of existing rule-based,externalinformation-based,human-based information error detection and correction techniques for relational databases to a great extent.The advantages of Web-based techniques include less dependence on the sufficiency of the database,more styles of constrains,wider applicability and more accurate repairs.In this review,we detail the advantages and challenges of Web-based information error detection and correction techniques.We propose a technological framework and believe it should include four components,including Web-based information expansion model,Web-based error detection algorithms,Web-based error correction algorithms and Web-based evaluation models for error detection and correction algorithms.Based on the framework,we comprehensively review current research works on the topics like Web-based error detection techniques,Web-based error correction techniques and Web-based information expansion techniques.We also refine out two key scientific problems which all Web-based information error detection and correction techniques must concern.Furthermore we prospect some future research topics and ideas.%信息质量已经成为诸多应用领域所面临的一个重要问题,自动检测和修复信息系统中的信息错误是改善信息质量的有效手段.利用Web对关系数据库中的信息进行扩展以助于信息错误的自动检测与修复具有对待检测与修复的信息本身依赖少、信息质量规则更灵活、适用性更广以及信息修复相对更准确等优势,可以有效克服现有的基于规则、基于扩展信息和基于人机交互的信息错误检测与修复技术的不足.文中详细分析了基于Web信息的信息错误自动检测与修复技术的优势及所面临的挑战,提出了基于Web信息的信息错误自动检测与修复技术框架.该框架包括:Web信息自动拓展模型、基于Web信息的信息错误自动检测算法、基于Web信息的信息错误自动修复算法和基于Web信息的信息错误自动检测与修复算法的可靠性评估模型.基于上述框架,文中系统总结了基于Web信息的信息错误自动检测技术、信息错误自动修复技术以及信息扩展核心技术三个方面的研究进展,提炼出了基于Web信息的信息错误自动检测与修复技术需要解决的关键科学问题,对未来的研究方向进行了展望并且讨论了初步的研究思路.