基于AR-LSTM-BP的CPI组合预测模型
The CPI Combination Prediction Model Based on AR-LSTM-BP
孙春 1庄科俊 1崔培贤1
作者信息
- 1. 安徽财经大学 统计与应用数学学院,安徽 蚌埠 233041
- 折叠
摘要
针对居民消费价格指数(CPI)预测准确性的问题,提出一种AR-LSTM-BP组合预测模型.首先分别用回归(AR)、长短时记忆网络(LSTM)和BP神经网络这三种模型对CPI预测,并对预测结果进行比较分析;随后引入诱导有序加权调和平均算子(IOWHA)的概念,构建AR-LSTM-BP组合预测模型.结果表明,IOWHA组合预测模型的误差均小于单项预测模型,预测结果准确性较高,能够更好地反映CPI的波动走势.
Abstract
This paper proposes an AR-LSTM-BP combination prediction model to address the accuracy of predicting consumer price index(CPI).First,we select three models to predict the CPI:the AR model,the LSTM model and the BP neural network model.Then we introduce the concept of induced ordered weighted harmonic mean operator(IOWHA),and construct the combination prediction model.The results show the IOWHA combination prediction model performs at a better prediction level than other individual prediction models,and the accuracy of the prediction results is high,which can reflect the fluctuation trend of CPI quite well.
关键词
CPI/组合预测模型/自回归模型/IOWHA算子Key words
CPI/AR model/IOWHA operator/combination prediction model引用本文复制引用
基金项目
安徽省高校科研计划重点项目(2022AH050565)
出版年
2024
NETL
NSTL
万方数据