基于CNN-LSTM的原煤产量预测模型
Raw coal production prediction model based on CNN-LSTM
张天宇 1王淼馨 1张正和 1王泽霖 1刘海涛2
作者信息
- 1. 辽宁工程技术大学安全科学与工程学院,辽宁葫芦岛 125105
- 2. 辽宁工程技术大学理学院,辽宁阜新 123000
- 折叠
摘要
为了准确地预测原煤产量,本文选择结合卷积神经网络(convolutional neural network,CNN)与长短期记忆网络(long short-term memory,L-STM)算法,建立基于 CNN-LSTM 的原煤产量预测模型.使用 2010年 1 月至 2021 年 12 月中国原煤产量的月度数据作为训练集,2022 年 1 月至 2022 年 12 月的数据作为检验集.利用训练后的模型预测 2023 年 1 月至 2023 年 12 月的原煤产量.通过与其他两种单一模型进行对比,并根据绝对相对误差评估模型预测结果.结果表明:CNN-LSTM 模型的原煤产量预测结果与实际值的绝对最大误差为 4.98%,预测精度显著提高,得到了 2023 年一整年的月度原煤产量预测结果,为国家未来发展和企业规划提供了科学的指导依据.
Abstract
To accurately predict raw coal production,this paper chooses to combine convolutional neural network(CNN)and long short-term memory(LSTM)algorithms to establish a raw coal production prediction model based on CNN-LSTM.Monthly data on China's raw coal production from January 2010 to December 2021 are used as the training set,and data from January 2022 to December 2022 are used as the test set.The raw coal production from January 2023 to December 2023 was predicted using the trained model.The model predictions are evaluated by comparing them with two other single models and based on the absolute relative error.The results show that the absolute maximum error between the raw coal production prediction results of the CNN-LSTM model and the actual value is 4.98%,the prediction accuracy is significantly improved,and the monthly raw coal production prediction results for the whole year of 2023 are obtained,which provides a scientific guiding basis for the future development of the country and enterprise planning.
关键词
安全工程/原煤产量预测/CNN-LSTM/绝对相对误差/预测精度Key words
safety engineering/forecast of raw coal production/CNN-LSTM/absolute relative error/prediction accuracy引用本文复制引用
出版年
2024
NETL
NSTL
万方数据