大数据近似分析方法综述

A review of big data approximate analytics

大数据分析旨在从大量复杂的数据中获取价值。查询驱动的数据分析是大数据分析中最主要的部分。由于数据量的庞大,在大数据上获取准确的分析结果将带来极大的存储和计算代价。为解决这一困难,大数据近似分析方法应运而生。本文将主要针对大数据近似分析中的频率估计问题、近似查询处理问题、查询选择性估计问题近十年的解决方法进行总结和归纳。不同于以往以数据库为主视角的分析方法的总结,本文中将涵盖近几年应用或结合机器学习方法来处理上述问题的新方法。

Big data analytics aims to obtain value from a large amount of complex data.Query-driven data analytics is the most important part of big data analytics.Due to the huge amount of data,obtaining accurate analysis results on big data will bring great storage and calculation costs.To solve this problem,big data approximate analysis methods came into being.This article mainly summarizes the frequency estimation methods,approximate query processing methods,and query selectivity estimation methods in big data analytics in the past ten years.Different from the previous summary of analysis methods based on the database perspective only,this article will cover the new methods of applying or combining machine learning methods to deal with the above problems in recent years.