摘要
海量用户需求数据的高效预判筛选,有助于海量数据下的快速查询所需数据。由于用户需求的不同,使得产生的用户需求数据多样化。传统筛选方法主要通过对用户需求数据进行预处理,分析其特征属性,实现数据筛选,忽略了预判在数据筛选中的重要性,导致筛选结果精度低的问题。提出基于时间序列的海量用户需求数据预判筛选方法,以用户需求数据的高效预判筛选原理为基础,通过统计学理论对用户需求数据进行预测,基于K-Means算法确定用户需求数据间距离,采用支持向量机方法结合回归分析,对用户需求数据进行预判。利用影响用户需求的各种不规则因子对用于需求数据进行时间序列分析,得到数据序列,并计算用户需求数据个属性值,获得数据筛选各项权重值,完成海量用户需求数据高效预判筛选。仿真结果表明,采用上述方法进行用户需求数据预判筛选,其预判及筛选结果精度要优于传统方法,具有一定的优势。
The efficient prejudgment and screening of massive user demand data are conducive to the quick query of demand data in massive data.Due to the variety of data,the traditional method mainly analyzes the characteristic attribute,but ignores the importance of prejudgment in the data screening,leading to the low precision of result.In this paper,a method of prejudging and screening massive user demand data based on time series was proposed.Based on the efficient prejudgment and screening principle of user demand data,the user demand data was predicted by statistical theory.Based on K-Means algorithm,the distance between user demand data was determined.Moreover,the support vector machine method was combined with regression analysis to prejudge user demand data.In addition,various irregular factors influencing the user demand were used to analyze the time series of user demand data,so as to obtain the data sequence.Meanwhile,each attribute value of user demand data was calculated to obtain the weight value of data screening.Finally,efficient prejudgment and screening of massive user demand data was completed.Simulation results show that the proposed method for prejudging and screening user demand data is superior to the traditional method,which has some advantages.
作者
罗琼
林若钦
LUO Qiong;LIN Ruo-qin(South China Institute of Software Engineering,Guangzhou University,Guangzhou Guangdong 510990,China)
出处
《计算机仿真》
北大核心
2019年第12期374-377,共4页
Computer Simulation
关键词
用户需求
数据
高效
预判
筛选
User needs
Data
Efficient
Prejudgment
Screening