网络向量自回归模型的模型平均
Model Averaging for Network Vector Autoregressive Models
潘颖 1李婷婷1
作者信息
- 1. 西南大学数学与统计学院，重庆 400715
- 折叠
摘要
网络向量自回归(NVAR)模型是分析和评估社交网络和金融网络等高维时间序列数据的重要工具。本文首先提出了一种基于Mallows准则的模型平均方法,用于整合不同的NVAR模型的预测结果。其次,该方法所得到的权重通过理论证明是具有渐近最优性的。最后,通过数值分析和实证分析表明,相较于其他模型选择和模型平均方法,本文所提出的方法具有更好的表现效果。
Abstract
Review Article: The Network Vector Autoregressive (NVAR) model is an important tool for analyzing and evaluating high-dimensional time series data such as social networks and financial networks. This paper proposes a model averaging method based on the Mallows criterion to integrate the prediction results of different NVAR models. Furthermore, the study confirms that the weights obtained by the MMA method are asymptotically optimal. Finally, numerical analysis and empirical analysis demonstrate that the proposed method outperforms other model selection and model averaging approaches in terms of predictive performance.关键词
网络向量自回归模型/模型平均/Mallows准则Key words
Network Vector Autoregressive Model/Model Averaging/Mallows Criterion引用本文复制引用
潘颖,李婷婷.网络向量自回归模型的模型平均[EB/OL].(2026-05-26)[2026-05-27].http://www.paper.edu.cn/releasepaper/content/202605-121.学科分类
计算技术、计算机技术
