基于哈希技术的高维数据相似性搜索研究进展  

Research on High-dimensional Data Similarity Search on Hashing Technology

作　　者：胡会南[1] 

机构地区：[1]宁波大学信息科学与工程学院,宁波315211

出　　处：《无线通信技术》2017年第1期41-45,共5页Wireless Communication Technology

摘　　要：数据检索中的相似性搜索是数据处理的基础问题之一,被广泛应用于机器学习、模式识别、图像检索等领域。本文针对哈希函数构建过程的不同分为基于位置敏感哈希的哈希索引技术和基于位置敏感哈希与布隆过滤器结合的哈希索引技术两部分进行综述。总结了具有代表性的哈希算法的研究现状,分析比较了各算法性能的主要差异,最后探讨了该技术尚需进一步研究的若干问题。Similarity search in data retrieval is one of the basic problems in the fields of computer science.It is widely used in machine learning,pattern recognition,image retrieval and other fields.According to different hash function＇s construction,this paper began to explain it from two aspects：one was based on locality sensitive hashing and the other was based on locality sensitive hashing with bloom filter.We summarized representative hash algorithms.Then we analyzed the differences between the performances of them.Finally,we discussed some problems that needed to be further studied in Similarity search.

关 键 词：高维数据 相似性搜索 位置敏感哈希 布隆过滤器 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]