文章摘要
邹鸿飞,王建州.一种基于差分灰狼算法的消费者信心预测指数的设计[J].数量经济技术经济研究,2019,(2):120-134
一种基于差分灰狼算法的消费者信心预测指数的设计
Design of a Consumer Confidence Forecasting Index Based on Differential Grey Wolf Algorithm
  
DOI：
中文关键词: 消费者信心指数  混合神经网络  预测模型  差分灰狼优化算法
英文关键词: Consumer Confidence Index  Hybrid Neural Network  Prediction Model  Differential Grey Wolf Optimization Algorithm
基金项目:本文获得国家社会科学基金重大项目“大数据时代雾霾污染经济损失评估及防治对策研究”（17ZDA093）的资助。
作者单位
邹鸿飞 东北财经大学统计学院 
王建州 东北财经大学统计学院 
中文摘要:
      研究目标：在大数据和互联网经济发展的背景下,有效预测消费者信心指数(CCI)以保证相关政策的制定。研究方法：基于完全集合经验模态分解(CEEMD)-差分灰狼算法(DEGWO)-BP神经网络(BPNN)，建立消费者信心指数预测模型，并运用DM检验法对该模型与对比模型的预测性能进行测试。研究发现：引入CEEMD法能够有效解决误差序列随机性强等缺陷;新提出的预测+模型较对比模型的预测精度明显提高，泛化能力有所增强，且更能够精准捕捉CCI的变化规律。研究创新：将CEEMD-DEGWO-BPNN模型应用于CCI预测中。研究价值：新提出的组合预测模型能够为CCI预测提供新方法，且有效地提高预测精度。
英文摘要:
      Research Objectives: In the context of big data and Internet economic development, the Consumer Confidence Index (CCI) is effectively predicted to ensure the formulation of relevant policies. Research Methods: In this paper, based on the complete set empirical mode decomposition (CEEMD)-differential grey wolf algorithm (DEGWO)-BP neural network (BPNN), the consumer confidence index prediction model is established, and the prediction performance of the model and the comparison model is tested by DM test. Research Findings: The introduction of CEEMD method can effectively solve the defects of strong randomness of error series; the proposed prediction model has higher prediction accuracy than the comparison model, and has strong generalization ability, which can capture the variation law of CCI. Research Innovations:The CEEMD-DEGWO-BPNN model is applied to CCI prediction. Research Value:The proposed combined prediction model can provide a new method for CCI prediction and improve prediction accuracy.
查看全文