基于改进BigFIM算法的网络信息大数据高频数据项挖掘算法研究  

Research on Frequent Data Item Mining based on Improved BigFIM in Network Information Big Data

作　　者：罗莉[1] 

机构地区：[1]四川交通职业技术学院,四川成都611130

出　　处：《激光杂志》2016年第7期135-140,共6页Laser Journal

基　　金：国家自然科学基金(61573081)

摘　　要：超大量的数据从诸如传感器、社交媒体、互联网应用等物联网产生,这些数据被统称为大数据。传统的工具和技术无法处理大数据。为了从大量数据中提取出对新技术有益的信息,大数据的挖掘尤为重要。非常受关注的关联规则挖掘和高频数据项挖掘都需要在内存中调入全部数据,但是大量的数据并不适合内存存储。为解决这一难题,业界提出了MapReduce进行并行的大数据处理。本文提出的改进Big FIM算法(Improved Big FIM,IBFIM)运行于MapReduce架构下用于大数据的挖掘.IBFIM相对于Big FIM增加了对更大规模数据的支持并提高了数据挖掘的速度。该研究为更快速、更高效地并行挖掘大数据内容提供参考。Big data is regarded as the huge amount of data generated from sensors, social networks, internet apps, etc. Traditional tools and techniques are incapable Of handling big data. To extract useful information from big data, data mining becomes essential. Both correlation rule mining and frequent data item mining require full load of dataset in main memory. However, huge amount of data is unsuitable for memory storage. To solve this problem, researchers proposed MapReduce to handle bigdata in parallel mode. An Improved BigFIM （IBFIM） is proposed in this paper. IBFIM runs on MapReduce framework to mine big data. Compared with BigFIM, IBFIM adds the ability to enhance the time consumption and support to scale data. IBFIM could be referred when considering faster and more efficiently parallel big data mining.

关 键 词：大数据 优化控制 加速算法 

分 类 号：TN911[电子电信—通信与信息系统]