Online Publishing Date:2018-08-20 16:18

基于人工智能技术的大数据分析方法研究进展
Progress of big data analytics methods based on artificial intelligence technology

人工智能、大数据、云计算、物联网等信息技术为推动集成制造快速发展提供了关键技术手段。近年来,采用人工智能技术进行大数据分析取得了突破性进展。系统总结了基于人工智能技术的大数据分析方法的最新研究进展。从大数据的聚类、关联分析、分类和预测4个主要的数据挖掘任务出发,分析了大数据环境下机器学习的研究现状;针对深度学习这一热点,总结了基于MapReduce、Spark的分布式深度学习实现,以及面向大数据分析的深度学习算法改进相关研究;从群智能、进化算法两方面梳理了基于计算智能的大数据分析相关研究;针对大数据平台,特别对大数据分析和深度学习集成框架进行了归纳,介绍了大数据机器学习系统和算法库;分析了大数据分析中人工智能技术面临的主要挑战,并提出了进一步的研究方向。

Artificial intelligence,big data,cloud computing,Internet of things and other information technologies promote the development of integrated manufacturing.Remarkable achievements were achieved in the methods of big data analytics with artificial intelligence technology.The latest research progress of big data analytics methods based on artificial intelligence was summarized comprehensively.A summary of research on machine learning was respectively introduced at first,including big data clustering,correlation analysis,classification and prediction.For the deep learning,a hotspot of research in machine learning,distributed deep learning models based on MapReduce/Spark and other improved deep learning algorithms for big data were discussed especially.The big data analytics based on computational intelligence were discussed from swarm intelligence and evolutionary algorithms two aspects.Furthermore,the engineering implementation of distributed computation platforms for big data were described,including the integrated frameworks for the distributive deep learning,big data machine learning systems and algorithms library.The challenges and the possible research directions of artificial intelligence technologies for big data analytics were put forward.

国家自然科学基金资助项目(61873240; 61572438; 61702456)~~;

大数据; 人工智能; 机器学习; 深度学习; 计算智能;

big data; artificial intelligence; machine learning; deep learning; computational intelligence;

10.13196/j.cims.2019.03.001

TP311.13;TP18

2404529-54719956K