基于Prophet-LSTM模型的柑橘价格预测研究
Citrus price prediction based on Prophet-LSTM model
黄楚欣 1吴小玉 1李菁琳 1董跃飞1
作者信息
- 1. 华南农业大学数学与信息学院信息工程系,广州 510642
- 折叠
摘要
鉴于柑橘价格的波动对市场和农民收益具有显著影响,因而对柑橘价格预测的研究不仅必要,且具有重要的意义.为准确预测柑橘价格,根据全国农产品商务信息公共服务平台 2018年1月1日至2022年9月24日的芦柑数据,分别使用Prophet模型、LSTM模型、Prophet-LSTM相结合的模型对柑橘价格进行建模和预测.研究结果显示Prophet-LSTM模型的MAE和MAPE最低,表明Prophet-LSTM模型预测效果最佳、更具稳定性,适合于柑橘价格的预测.
Abstract
The fluctuation of citrus price has a significant impact on the market and farmers'income,so the research of citrus price prediction is of great significance.In order to accurately predict the price of citrus,the Prophet model,LSTM model,and the combined model of Prophet and LSTM were respectively used to model and forecast the price of citrus based on the data of the Na-tional Agricultural product Business information Public Service Platform from January 1st,2018 to September 24th,2022.The re-sults showed that the MAE values of Prophet model,LSTM model,and the combined model of Prophet and LSTM were 0.149,0.291 and 0.136,and the MAPE values were 0.031,0.056 and 0.028,respectively.The results show that Prophet-LSTM model has the best prediction effect and is more stable,which is suitable for the prediction of citrus price.
关键词
柑橘/价格预测/Prophet-LSTM/时间序列/经济作物Key words
citrus/price forecasting/Prophet-LSTM/time series/cash crop引用本文复制引用
基金项目
2022年华南农业大学大学生创新创业训练计划项目(S202210564125)
出版年
2024
NETL
NSTL
万方数据