数值信息抽取研究概述及应用分析

A Overview of Numerical Information Extraction Research and Application Analysis

【目的/意义】面对海量的信息,人们需要更为高效准确的信息获取方式。数值信息抽取的研究使隐含在无序信息载体中的大量有价值数值信息可以得以利用,从而满足科研工作者数据驱动型研究的信息需求。【方法/过程】本文旨在总结和归纳数值信息抽取研究的相关内容,包括数值信息抽取的内涵、数值信息抽取研究概况、面临的困境和制约因素以及应用等。【结果/结论】数值信息抽取仍然面临着巨大的挑战,且现有的数值信息抽取研究较少,对于数值信息的抽取,基于规则和统计学习的方法各有利弊,总体来说,基于规则的抽取方法仍是主流方法。

【Purpose/significance】Faced with massive amounts of information, people need more efficient and accurate ways to obtain information. There are also a lot of useful disordered information,research on numerical information extraction can help us use this information to meet the information needs of scientific research workers who does data-driven research. 【Method/process】This paper aims to summarize the relevant content of numerical information extraction research,including the connotation of numerical information extraction, the research overview of numerical information extraction,the dilemmas, the main constraints and applications.【Result/conclusion】The extraction of numerical information still facesenormous challenges, and there is less research on the extraction of existing numerical information. For the extraction of numerical information, methods based on rules and statistical learning have advantages and disadvantages. Overall,rules-based extraction methods are still the mainstream method.