基于异构数据源的政法信息共享平台数据预处理系统研究

摘要

针对分散在公安、检察院、法院、司法部门的没有信息共享平台的现状,提出了一个基于异构数据源的政法网平台数据预处理系统.设计采用了在不改变政法系统架构的基础上搭建政法信息共享平台,以达到信息互通、资源共享.系统应用的实验结果表明:基于规则库的多级数据预处理算法通过连续的样本训练建立越来越完善的规则库,不断提高后续数据抽取质量,并通过上述多级规则库匹配预处理策略,将其分步嵌入到数据应用系统中,除非常少的个别错误数据要单独汇总处理外,脏数据输出较少,预处理速度也比较适中;基于规则库的数据预处理算法达到了政法信息共享平台的建设目标,使政法业务信息在各职能部门能相互共享,且为领导决策提供及时可靠的依据.%Based on scattered in public security, procuratorates, courts and judicial departments have no the present situation of the information sharing platform, proposed a network platform for data preprocessing of politics and law system based on heterogeneous data sources. Design adopted without changing of politics and law system framework built on the basis of the information sharing platform of political science and law, in order to achieve information exchange, resource sharing. System application of the experimental results show that the multilevel data pretreatment algorithm based on rule base through continuous sample training set up more and more perfect the rules of the library, and constantly improve the quality of the follow-up data extraction, and through the multi-stage rule base matching pretreatment strategy, the embedded in the data application system step by step, except the data on very few individual mistakes to separate aggregate processing, dirty data output is less, pretreatment are more moderate speed;Data pretreatment algorithm based on rule base is achieving the target of information sharing platform construction of politics and law, make the business information of politics and law in various functional departments can share each other, and provide timely and reliable basis for leadership decision-making.