基于模型遴选规则的汇率自适应组合预测  

Combinational Forecasting of Self-Adaptive Exchange Rate Based on Rules of Selection Models

作　　者：蒋传进[1] 

机构地区：[1]上海商学院信息与计算机学院

出　　处：《统计与决策》2017年第16期74-78,共5页Statistics & Decision

基　　金：国家自然科学基金面上项目(61571326);国家自然科学基金青年项目(61201179);上海自然科学基金青年项目(13ZR1429600)

摘　　要：文章阐述了组合预测模型遴选规则的产生机理,对此遴选规则的具体形式以及组建方式进行了研究分析。为了使构建于模型遴选规则基础之上的组合预测模型更具鲁棒性及更广的适用领域,通过基于NARX神经网络的自适应调节机制来提升遴选规则的学习能力,提高了组合预测的精度和稳定性。最后以一个汇率预测的实证分析证明了该模型的有效性。This paper expounds the generating mechanism of selection rules of the combinational forecasting models, and analyzes the specific form of the selection rules and the forming method. In order to make the combinational prediction models constructed on the model selection rules more robust and applicable, this paper introduces a self-adaptive regulation mechanism based on NARX neural network to improve the learning ability of the selection rules, thus improving the precision and stability of combinational forecasting. Finally, the paper takes an exchange rate as an empirical analysis to verify the effectiveness of the proposed model.

关 键 词：组合预测 自适应调节 NARX神经网络 汇率预测 

分 类 号：F224[经济管理—国民经济] F832