SGER: Investigating Research Problems Related to Exchanging Information from Heterogenous Data Sources

这项探索性研究（SGER）项目的小额赠款将与德克萨斯州的两个州机构和德克萨斯州盖恩斯维尔市合作，就传统刑事司法数据库的主题扩展数据集成技术。从技术上讲，该项目将研究(1)构建中介数据库的体系结构解决方案，以及(2)自动化模式匹配的新方法。更具体地说，研究将演示将虚拟全局数据库和各种数据源之间的映射表示为XQuery查询的可行性。XQuery应该提供处理数据的效率、描述数据映射的能力以及改进标准化的潜力。该项目的第二个任务将解决异构数据库之间的模式匹配问题，研究基于计算语言学中已经成功应用的技术的几个过程。本研究是关于元数据开发和基于语言的模式匹配自动化使用相关的集成信息技术。拟议研究的工作将垂直整合，将大量的精力用于研究拟议项目的基本原则，开发基于研究的实用技术，以及说明技术的演示和测试平台。拟议研究的主要目标将是创建一个原型信息系统，用于整合德克萨斯州境内遗留的刑事司法信息。然后将评估原型的有效性，评估结果将用于开发数据同化和模式集成的通用方法。拟议的研究将通过帮助各县将遗留的刑事司法数据传输到该州的信息系统而使德克萨斯州受益。在研究结束时，团队将完成：(1)支持司法信息数据集成的原型架构和一套规范；(2)一套用于自动模式匹配的工具和技术。除了研究生直接参与论文研究之外，该项目还将为数据库、自然语言处理和机器学习等领域的课程开发人工制品。这项研究将涉及本科生和附近社区的成员。

This Small Grant for Exploratory Research (SGER) project will extend data integration technologies in partnership with two Texas State agencies and the City of Gainesville TX on the topic of legacy criminal justice databases.Technically the project will examine (1) architectural solutions to constructing mediated databases, and (2) new methods for automating schema matching. More specifically,the research will demonstrate the feasibility of representing the mappings between the virtual global database and the various data sources as XQuery queries. XQuery should provide efficiencies in processing the data, ability to describe the data mappings, and potential for improving standardization. The project's second task will address the schema matching problem among heterogeneous databases, investigating severalprocedures based on techniques already successfully employed in Computational Linguistics.Intellectual Merit This research is on integrated information technology related to the development of metadata and the use of automation of schema matching by means of linguistic based techniques. The work of the proposed study will be vertically integrated, with a significant amount of effort devoted to research on the principles underlying the proposed projects, development of practical technology based on the research, and demonstrations and test beds that illustrate the technology. The major goal of the proposed research will be the creation of a prototype information system for integrating legacy criminal justice information within the State of Texas. The effectiveness of the prototype will then be assessed and results of the evaluations will be used to develop generalpurpose approaches to data assimilation and schema integration.Broader Impact The proposed research will benefit the State of Texas by helping counties transmit legacy criminal justice data to the State's information system. By the end of the proposed research, the team will complete: (1) a prototype architecture and a set of specifications that support data integration of justice information; and (2) a set of tools and techniques for automated schema matching. Beyond the direct involvement of graduate students conducting research towards their theses and dissertations, the project will develop artifacts for courses in the areas of database, natural language processing, and machine learning. The research will involve undergraduate students and members of the nearby community.