Flexible and Efficient Text Storage, Management, and Retrieval

点击翻译按钮获取中文摘要

Applications relying on text must be built upon solid foundations of data organization similar to those that underlie conventional database systems. Thus the central long-term objective of the research program is to support document storage and management by applying sound database principles to this domain of information. The challenge is to discover how the complexity of text, with its intricate structure and diversity of expression, can be efficiently and effectively accessed and managed. In the short term, this objective will be addressed through the design and implementation of improved search engines. The research program will include three related investigations. I will explore problems of general-purpose search engines, including aspects specific to reference texts, where nested structure is central to locating data. For example, I will investigate how the mechanism of database views can be best applied to describe, design, and implement an index that provides efficient search, including support for phrasal queries. I will also examine aspects specific to providing optometry researchers with access to Electronic Medical Records, where free-text fields include many spelling errors and where ensuring patient privacy is mandatory. In the third thrust, I will concentrate on aspects specific to mathematics information retrieval, where the relative positioning of symbols is at least as important as the particular symbols used. The inadequacy of existing solutions leads to the design of improved algorithms, which I then analyze, implement, and evaluate. Standard benchmarks provide the basis for comparative evaluations, but often additional test collections must be developed to examine specific aspects of the problems from unexplored perspectives. Graduate students will participate in all aspects of this research, resulting in HQP well-prepared to contribute to this area. The centrality of documents in recording and preserving information suggests that this research remains timely and crucial to economic growth in Canada. The integration of information retrieval and database management is well-recognized as being important to Canadian business and could be equally important in promoting Canadian culture. More specifically, the reliance on text search has become a universal need, with search engines providing critical roles in gathering information internet-wide, from across one or more enterprises, or from one’s own personal data collections. In addition, other text applications are found throughout business and government. Publishers, data providers (e.g., via the World Wide Web), and organizations that rely on any form of text-dominated knowledge base for conducting their internal and external business will benefit from specific tools that arise from this research as well as from the theory, which will provide a framework for designing their text management systems.