广义洛特卡分布特征参数解区间波动研究  

Study on Fluctuation of Solution Interval about Generalized Lotka Distribution Parameter

作　　者：陈道兰[1] 毛一波 Chen Daolan;Mao Yibo(Library of Chongqing University of Arts and Sciences, Chongqing 402160;Institute of Mathematies and Eeonomics, Chongqing University of Arts and Seiences, Chongqing 402160)

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：《情报探索》2018年第5期21-25,共5页Information Research

摘　　要：[目的/意义]研究在数据采样不完全情形下广义洛特卡分布特征参数解区间的波动情况。[方法/过程]以《情报理论与实践》期刊为例,对完整数据随机采样得到非完全采样的作者信息,计算广义洛特卡分布特征参数的解区间,描述解区间受数据采样不全引起的波动情况。[结果/结论]数据采样不全将影响解区间的范围,但二者的关系又不完全是负相关关系。数据采样的完整程度达到70%以上可使解区间的中心的波动幅度控制在5%以内,而当数据采样的完整程度达到80%以上,解区间的半径波动幅度能勉强小于20%。为使解区间的估计更接近真实值,应尽可能地对数据进行完全采样。[Purpose/significance]The paper is to study on the fluctuation of solution interval about generalized Lotka distribution parameter under incomplete sampled data. [Method/process]The paper takes Journal of Information Theory and Application（ITA） for example, gets incomplete sampling author information by random sampling of complete data, calculates solution interval about generalized Lotka distribution parameter, describes the fluctuation of solution interval because of incomplete data sampling. [Result/conclusion]Incomplete data sampling will affect the solution interval, but the influence is not entirely negative correlation. When the integrity of data sampling is more than 70%, the fluctuation range of the solution interval center can be controlled within 5%, while the integrity of data sampling reaches more than 80%, the radius fluctuation range of the solution interval can barely be less than 20%. The data should be counted as fully as possible in order to make the estimation of solution interval close to the true values.

关 键 词：非完全采样 洛特卡分布 特征参数 解区间 波动幅度 

分 类 号：G350[文化科学—情报学]