SBIR Phase I: Xtractica: A System for Extracting Coherent Data from Documents

SBIR 第一阶段：Xtractica：从文档中提取一致数据的系统

• 批准号: 0128508
• 负责人: Hasan Davulcu
• 金额: $ 9.98万
• 依托单位: XSB, INC.
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2002
• 资助国家: 美国
• 起止时间: 2002-01-01 至 2002-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0128508&HistoricalAwards=false
• 关键词: SBIR; Phase; Xtractica; System; Extracting

This Small Business Innovation Research (SBIR) Phase I project will investigate the feasibility of designing and building a software system: Xtractica. This software system will allow domain experts to specify programs that transform unstructured or partially structured data from a variety of document sources, such as World Wide Web sites, PDF files and text into structured, coherent and readily usable information. Xtractica will consist of a set of tightly integrated powerful syntactic and semantics-driven data extraction technologies that are managed from a graphical user interface to retrieve information that was created for human understandability, and extract and reason about it to create knowledge that can support automated decision making and transactions. An important feature of Xtractica is that users can rapidly create extractors by simply supplying examples of the data to be extracted. Thus it will empower users who are knowledgeable about their application domains but are not necessarily trained as computing technologists, to structure data into knowledge.The Phase I project will develop the operational specifications of Xtractica and determine its feasibility by prototyping its critical components. Phase 2 will then produce a fully functional Xtractica system based on results from Phase I. Finally Phase 3 will make Xtractica commercially available to clients with diverse business interests including content aggregation, e-procurement, ERP and supply chain management vendors.

这个小企业创新研究（SBIR）第一阶段项目将研究设计和构建一个软件系统的可行性：Xtractica。这一软件系统将使领域专家能够指定程序，将来自万维网网址、PDF文件和文本等各种文件来源的非结构化或部分结构化数据转换为结构化、连贯和易于使用的信息。Xtractica将由一组紧密集成的强大的语法和语义驱动的数据提取技术组成，这些技术从图形用户界面进行管理，以检索为人类理解而创建的信息，并提取和推理这些信息，以创建可以支持自动决策和事务的知识。Xtractica的一个重要特性是用户可以通过简单地提供要提取的数据的示例来快速创建提取器。因此，它将使那些熟悉其应用领域但不一定受过计算机技术培训的用户能够将数据结构化为知识。第一阶段项目将制定Xtractica的操作规格，并通过制作其关键组件的原型来确定其可行性。第二阶段将根据第一阶段的结果生产一个功能齐全的Xtractica系统。最后，第三阶段将使Xtractica商业化，提供给具有不同商业利益的客户，包括内容聚合、电子采购、ERP和供应链管理供应商。