基于非结构化文本检索模型综述

发布时间：2018-03-22 06:04

本文选题：相似度计算　切入点：检索模型　出处：《计算机应用研究》2017年06期 　论文类型：期刊论文

【摘要】：随办公信息化、生活网络化不断推进,诸如企业产品问题描述、Web用户评论、通信文本信息等新生的非结构化文本数据也伴随着快速的增长以及其不断积累。这对于如何能准确、高效地检索到用户真实需求的文本信息提出了新的要求和挑战。检索模型对检索准确度、效率等具有决定性影响。近年来,大量新兴方法融入到文本的检索模型中,使模型本身变得纷繁复杂,同时传统模型间的界限变得模糊。从非结构化文本数据的检索需求出发,归纳检索模型的定义和通用框架;进而基于检索词项相似性计算采用的数学理论,对检索模型进行分类,并详细阐述各类模型的发展脉络、分析其优缺点及适用场景。最后,讨论了新环境下海量文本检索模型面临的挑战及相关研究问题思考。
[Abstract]:With the development of office information and network, new unstructured text data, such as enterprise product problem description, Web user review, communication text information and so on, are growing rapidly and accumulating. In recent years, a large number of new methods have been integrated into the text retrieval model, which has a decisive effect on retrieval accuracy and efficiency. The model itself becomes complicated and the boundary between the traditional models becomes blurred. Based on the retrieval requirements of unstructured text data, the definition and general framework of the retrieval model are summarized. Then, based on the mathematical theory of similarity calculation, the retrieval model is classified, and the development of each model is described in detail, and its advantages and disadvantages and applicable scenarios are analyzed. This paper discusses the challenge of massive text retrieval model in the new environment and some related research issues.
【作者单位】：桂林电子科技大学广西云计算与大数据协同创新中心;桂林电子科技大学广西可信软件重点实验室;桂林电子科技大学广西自动检测技术与仪器重点实验室;
【基金】：国家自然科学基金资助项目(U1501252,61462017,61363005) 广西区自然科学基金资助项目(2014GXNSFAA118353,2014GXNSFAA118390,2014GXNSFDA118036) 广西自动检测技术与仪器重点实验室基金资助项目(YQ15110) 广西高等学校高水平创新团队及卓越学者计划资助项目
【分类号】：TP391.3

【相似文献】