摘要
【目的】在大数据环境下,寻找最佳的输入数据分片数量,以便改进统计算法的运行效率。【方法】以车牌统计为例,调整相关多个参数以改变算法中输入数据的分片数量,并分析不同参数下算法运行时间的差异。【结果】最佳的分片数量下的运行效率近似于最差的分片数量下运行效率的2倍。【结论】大数据处理中,合理的输入数据的分片数量有助于提高算法的运行效率。同时也分析了分片数量与算法运行时间的函数关系,以期找到最佳的分片数量区间。
[Purposes]In the big data environment,finding the optimal number of input data fragments is an effective means to improve the efficiency of statistical algorithms.[Methods]Taking the license plate statistics as an example,several related parameters(minsize,maxsize)were adjusted to change the number of input data fragments in the algorithm,and the different running time of the algorithm was analyzed under different parameters.[Findings]The running efficiency at the optimal number of fragments was found 1 times faster than that the worst number of fragments.[Conclusions]In the big data processing,the running efficiency of the algorithm is improved by the reasonable number of input data fragments.The functional relationship between the number of fragments and the running time of the algorithm was also analyzed to find the optimal quantity interval of fragments.
作者
汤星
范永胜
冯骥
钟贞
孔亚迪
TANG Xing;FAN Yongsheng;FENG Ji;ZHONG Zhen;KONG Yadi(College of Computer and Information Science,Chongqing Normal University,Chongqing 401331;College of Computer Technology and Engineering,Nanjing University of Science and Technologyt Nanjing 210094,China)
出处
《重庆师范大学学报（自然科学版）》
CAS
北大核心
2019年第6期98-103,共6页
Journal of Chongqing Normal University:Natural Science
基金
重庆师范大学（人才引进/博士启动）基金项目（No.17XCB008）
教育部人文社会科学研究项目（No.18XJC880002）
重庆市教育委员会科技项目（No.KJQN201800539）
关键词
车牌统计算法
分片数量
算法运行时间
license plate statistics algorithm
number of fragments
running time of algorithm
