摘要
提出一种基于概率统计的大数据分类优化方法.首先构建大数据分类的二项-泊松模型,在无穷维向量空间对分类目标函数进行概率密度估计,在数据聚类中心的几何邻域构建置信区间,采用置信度特征求解大数据分类的二项-泊松微分方程,然后以方程的稳定解作为聚类中心特征值,在Bernoulli空间通过概率统计实现对大数据的准确分类.
A large data classification method is proposed based on probability statistics.Firstly twoPoisson model construction of data classification is built,probability density functional of classification objective function is established in the infinite dimensional vector space,and confidence intervals are constructed for geometric neighborhood clustering in the data center.Two Poisson equation is solved by confidence discriminate in data classification and the characteristics of stable solution of equations is taken as the clustering center value to achieve accurate data classification in Bernoulli space.
作者
杨晓波
YANG Xiao-bo(Sichuan Information Technology College ,Guangyuan 628017, Sichuan, China)
出处
《内蒙古师范大学学报（自然科学汉文版）》
CAS
北大核心
2017年第2期174-177,共4页
Journal of Inner Mongolia Normal University(Natural Science Edition)
基金
四川省教育厅自然科学基金项目(14ZB0442)
关键词
大数据分类
概率统计
泛函
微分方程
large data classification
probability and statistics
functional
differential equation
