统计数据缺失对广义洛特卡分布特征参数取值影响的研究  

Research on the Influence of Data Missing on Characteristic Parameters of Generalized Lotka Distribution

作　　者：陈道兰[1] 毛一波 CHEN Daolan;MAO Yibo(Library, Chongqing University of Arts and Science, Chongqing 402160, China;Institute of Mathematics and Economics, Chongqing University of Arts and Sciences, Chongqing 402160, China)

机构地区：[1]重庆文理学院图书馆,重庆402160 [2]重庆文理学院数学与财经学院,重庆402160

出　　处：《农业图书情报学刊》2018年第6期166-170,共5页Journal of Library and Information Sciences in Agriculture

摘　　要：以《情报科学》期刊为例,通过随机构造含有缺失数据的作者信息,研究了在统计数据缺失的情况下广义洛特卡分布特征参数的变化情况。实验结果表明,广义洛特卡分布规律具有普适性,广义洛特卡分布特征参数的取值会受数据缺失的影响,但并不完全是正相关关系;数据缺失比例控制在20%以内可使特征参数值受影响的程度低于5%;为了使广义洛特卡分布特征参数的估计更接近真实值,所统计的数据应尽可能完全。Taking the Journal of Information Science as an example, this paper studied the changes of characteristic parameters of generalized Lotka distribution by randomly constructing missed author information. The experimental results showed that generalized Lotka distribution was universal, while the data missing affected characteristic parameters of generalized Lotka distribution, and the influences were not entirely positive correlation. The proportion of data missing within 20% can make the affected degree of characteristic parameters lower than 5%. The statistical data should be counted as fully as possible to make the estimation of generalized Lotka distribution parameters close to the true values.

关 键 词：数据缺失 洛特卡分布 特征参数 受影响程度 

分 类 号：G350[文化科学—情报学]