摘要
随着全球贸易的快速发展,商品价格的动态波动对经济活动和企业决策产生重要影响。为解决传统预测模型在处理时间序列非线性和高维数据时的局限性,研究提出了一种结合随机森林、鲸鱼优化算法和长短期记忆网络的价格动态波动预测模型。模型通过随机森林筛选关键特征,通过鲸鱼优化算法优化长短期记忆网络超参数,并基于优化的长短期记忆网络实现价格预测。实验结果表明,基于时间序列预测算法的贸易商品价格动态波动预测模型的预测准确率达到92.5%,均方根误差仅为0.15,优于传统模型。研究结果证明,随机森林与鲸鱼优化算法结合显著提升了长短期记忆网络模型的预测精度与稳定性,为复杂时间序列预测提供了有效方法。
With the rapid development of global trade,the dynamic fluctuations in commodity prices have a significant impact on economic activities and corporate decision-making.To address the limitations of traditional prediction models in handling non-linear and high-dimensional time series data,a price dynamic fluctuation prediction model combining random forest,whale optimization algorithm,and long short-term memory network is proposed.The model selects key features through random forest,optimizes the hyperparameters of the long short-term memory network through whale optimization algorithm,and implements price prediction based on the optimized long short-term memory network.The experimental results show that the prediction accuracy of the trade commodity price dynamic fluctuation prediction model based on time series prediction algorithm reaches 92.5%,with a root mean square error of only 0.15,which is better than traditional models.The research results demonstrate that the combination of random forest and whale optimization algorithm is significant.
作者
李亮亮
LI Liangliang(School of Economics and Trade,Anhui Business and Technology College,Hefei 231131,China)
出处
《佳木斯大学学报(自然科学版)》
2025年第7期155-159,共5页
Journal of Jiamusi University:Natural Science Edition
基金
安徽省高校教学研究项目(2023xjjy33)。
关键词
时间序列
贸易
商品价格
预测
长短期记忆网络
time series
trade
product price
forecast
long short term memory network
