区域经济发展潜力时间序列预测模型仿真
Simulation of Time-series Prediction Model of Regional Economic Development Potential
唐磊 1宋婷婷2
作者信息
- 1. 安徽工业经济职业技术学院 职业技能鉴定所,安徽 合肥 230051
- 2. 福州理工学院 计算与信息科学学院,福建 福州 350506
- 折叠
摘要
为提高区域经济发展潜力时间预测准确性,提出了区域经济发展潜力时间序列预测模型仿真.文章对区域经济发展数据归一化、降维重构;基于滑动窗口,引入决策树进行区域经济发展数据时序特征提取;利用自注意力机制线性映射函数进行区域经济发展数据特征融合;基于残差网络改进CNN网络,构建区域经济发展潜力时间序列预测模型.仿真结果表明该方法的预测拟合系数较高,平均绝对误差和均方根误差均低于0.4,提升了预测的准确性.
Abstract
In order to improve the accuracy of time prediction of regional economic development potential,time series prediction model of regional economic development potential is proposed.Normalization and reducing reconstruction of regional economic development data;introducing decision tree to extract temporal characteristics of regional economic development data;using the linear mapping function of attention mechanism;improving CNN network based on residual network to build the time series prediction model of regional economic development potential.The simulation results show that the prediction fit coefficient is high,and the mean absolute error and root mean square error are below 0.4,which improves the prediction accuracy.
关键词
时间序列/时序特征/区域经济Key words
time series/time sequence characteristics/regional economy引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据