基于多模态LSTM预测2023年中国粗钢表观需求量  

China’s apparent demand forecasting for crude steel in 2023 based on multitask LSTM

作　　者：崔煜 耿志华 丁东 CUI Yu;GENG Zhihua;DING Dong(Harbin Institute of Technology(Shenzhen),Shenzhen 518071,China)

机构地区：[1]哈尔滨工业大学(深圳),广东深圳518071

出　　处：《中国高新科技》2023年第17期57-59,共3页

摘　　要：中国粗钢表观需求量是衡量钢铁行业市场规模和发展趋势的重要指标之一。因此,准确预测中国粗钢表观需求量对于钢铁企业和政府决策者具有重要的意义。文章旨在研究中国粗钢表观需求量的预测问题,通过对中国粗钢市场的历史数据进行分析,结合宏观经济因素、行业特点和国际上其他国家的需求量数据,建立多模态LSTM预测模型。通过对比使用线性模型方案的预测效果,得出最优方案。该研究对于了解中国粗钢市场的发展趋势和制定相关政策具有一定的参考价值,所提出的预测模型和方案也可为其他行业的需求量预测提供参考和借鉴。The apparent demand for crude steel in China is one of the important indicators to measure the market size and development trend of the steel industry.Therefore,accurately predicting the apparent demand for crude steel in China is of great significance for steel companies and government.This article aims to study the problem of predicting the apparent demand for crude steel in China.By analyzing historical data of the Chinese crude steel market,combining macroeconomic factors,industry characteristics,and demand data from other countries internationally,a prediction model based on multitask LSTM is established.By comparing the predictive effects of using linear model schemes,the optimal scheme was obtained.This study has certain reference value for understanding the development trend of China’s crude steel market and formulating relevant policies,the prediction model and scheme proposed in this article can also provide reference and reference for demand forecasting in other industries.

关 键 词：神经网络 LSTM 多模态 钢材需求量 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]