多个关联确定性时间序列微分方程建模方法研究  

The Research of Differential Equations Modeling Methods of Correlative Deterministic Multivariate Time Series

作　　者：孙文凯[1] 胡晓华[1] 蒋文江[1] 

机构地区：[1]海南师范大学数学与统计学院,海南海口571158

出　　处：《海南师范大学学报（自然科学版）》2015年第1期9-14,共6页Journal of Hainan Normal University（Natural Science）

基　　金：国家自然科学基金资助项目(11361022)

摘　　要：从纯数学的角度,对多个关联确定性时间序列分别进行多次累加产生新序列,研究序列之间的关系,建立多元线性(或非线性)回归方程.给定显著性水平α,对每个回归方程进行显著性检验.在置信度1-α下建立微分方程组模型,从而揭示这些时间序列之间的关系,实现对原序列的预测和控制.最后用1995-2014年海南省GDP和接待旅游人数建立微分方程组模型并进行预测.From the perspective of pure mathematics, Deterministic multivariate time series were multiply accumulated to produce some new sequences, studying relationship among them and establishing the multiple linear or nonlinear regression model. If a significance level ot was given, the significance test for every regression equation was carded out. At confidence level 1-α, the differential equations models were established to reveal the relationship among all the time series, and fore- cast or control them. Finally, using the data for the value of GDP and the number of tourist reception in Hainan province from 1995 to 2014, differential equations model is established to forecast.

关 键 词：多个时间序列 累加 微分方程 预测 

分 类 号：O29[理学—应用数学]