[1]项晓宇,朱敏捷,周灵刚,等.基于机器学习的短期规上行业工业增加值预测[J].南京师大学报(自然科学版),2023,46(02):99-106.[doi:10.3969/j.issn.1001-4616.2023.02.013]
　Xiang Xiaoyu,Zhu Minjie,Zhou Linggang,et al.Short-Term Industrial Added Value Prediction of the Above-Scale Industry Based on Machine Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2023,46(02):99-106.[doi:10.3969/j.issn.1001-4616.2023.02.013]
点击复制

基于机器学习的短期规上行业工业增加值预测()
分享到：

《南京师大学报（自然科学版）》[ISSN:1001-4616/CN:32-1239/N]

卷:
第46卷
期数:
2023年02期
页码:
99-106
栏目:
计算机科学与技术
出版日期:
2023-06-15

文章信息/Info

Title:
Short-Term Industrial Added Value Prediction of the Above-Scale Industry Based on Machine Learning
文章编号:
1001-4616(2023)02-0099-08
作者:
项晓宇1; 朱敏捷1; 周灵刚1; 钟磊1; 闵富红2
(1.国家电网台州供电公司,浙江 台州 318000)
(2.南京师范大学南瑞电气与自动化学院,江苏 南京 210023)
Author(s):
Xiang Xiaoyu1; Zhu Minjie1; Zhou Linggang1; Zhong Lei1; Min Fuhong2
(1.Taizhou Electric Power Company,Taizhou 318000,China)
(2.School of Electrical and Automation Engineering,Nanjing Normal University,Nanjing 210023,China)
关键词:
规上工业; 工业增加值预测; 机器学习; Stacking算法
Keywords:
above-scale industry;  industrial added value prediction;  machine learning;  Stacking algorithm
分类号:
TP301.6
DOI:
10.3969/j.issn.1001-4616.2023.02.013
文献标志码:
A
摘要:
针对现有工业增加值预测存在经济数据统计滞后、单一模型预测精度差的问题,本文提出一种基于Stacking框架下短期规上行业的工业增加值预测模型,实现了预测时效性与精度的提升. 通过皮尔逊相关性系数检验,对浙江省某市4个重点规上行业的用电量、工业增加值进行分析,发现两者具有中强度的相关性,表明了基于行业用电量预测工业增加值方法的可行性. 接着,以随机森林算法、自适应增强算法、极致梯度增强算法3种模型作为基学习器,支持矢量回归机算法作为元学习器,搭建双层Stacking融合模型框架对规上行业用电量、工业增加值、当地气温数据进行模型训练测试,得到最终预测模型. 最后,将本文所提出的Stacking模型与单一模型预测误差指标进行实例对比分析,结果表明,该模型具有更高的预测精度,且采用月度收集的实时电力消费数据提升了预测时效性,能被更好地应用在“双碳”背景下工业增加值的预测场景中,也有利于政府分析经济发展趋势.
Abstract:
Considering the lagging indicators with the economic statistics and poor accuracy of the single model in the prediction of industrial added value, this paper proposes a short-term model(STM), via the Stacking algorithm, to improve the timeliness and accuracy of the prediction. With Random Forest Regressor, AdaBoost Regressor and Xgboost Regressor as the base learners and Support Vector Regression as the meta learners, STM is established in a two-layer stacking framework. Via the Pearson correlation analysis, the data of the electricity consumption, the local temperature, and the industrial added value of the above-scale industries in a city of Zhejiang province are investigated, the result of which shows that these data are quite related to each other. Therefore, via the training test of these data, STM can effectively predict the industrial added value through the industrial electricity consumption. The prediction results via STM are compared with that of the single model. The comparison shows that the monthly collected real-time electric consumption in Stacking model can improve the time efficiency of the prediction. Therefore, STM is quite suitable for the prediction of the industrial added value, and useful for the dual control system of the total carbon emissions and carbon emission intensity, which is helpful for the government to analyze the trend of economic development.

参考文献/References:

[1]刘自敏,邓明艳,杨丹,等. 降低企业用能成本可以提高能源效率与社会福利吗——基于交叉补贴视角的分析[J]. 中国工业经济,2020(3):100-118.
[2]顾光同,许冰. 中国工业增加值的半月预报:基于宏观月度数据[J]. 系统工程理论与实践,2018,38(8):1983-1993.
[3]DILEEP G. A survey on smart grid technologies and applications[J]. Renewable energy,2020,146(2):2589-2625.
[4]ZHU J M,LIU J P,WU P,et al. A novel decomposition-ensemble approach to crude oil price forecasting with evolution clustering and combined model[J]. International journal of machine learning and cybernetics,2019,10(12):3349-3362.
[5]MENG M,WANG L X,SHANG W. Decomposition and forecasting analysis of China's household electricity consumption using three-dimensional decomposition and hybrid trend extrapolation models[J]. Energy,2018,165(A):143-152.
[6]CONEJO A J,PLAZAS M A,ESPINOLA R,et al. Day-ahead electricity price forecasting using the wavelet transform and ARIMA models[J]. IEEE transactions on power systems,2021,20(2):1035-1042.
[7]CHOU J S,TRAN D S. Forecasting energy consumption time series using machine learning techniques based on usage patterns of residential householders[J]. Energy,2018,165(B):709-726.
[8]TAN M,YUAN S P,LI S H,et al. Ultra-short-term industrial power demand forecasting using LSTM based hybrid ensemble learning[J]. IEEE transactions on power systems,2020,35(4):2937-2948.
[9]RAMOS D,FARIA P,VALE Z,et al. Short time electricity consumption forecast in an industry facility[J]. IEEE transactions on industry applications,2022,58(1):123-130.
[10]CHEN K J,CHEN K L,WANG Q,et al. Short-term load forecasting with deep residual networks[J]. IEEE transactions on smart grid,2018,10(4):3943-3952.
[11]赵腾,王林童,张焰,等. 采用互信息与随机森林算法的用户用电关联因素辨识及用电量预测方法[J]. 中国电机工程学报,2016,36(3):604-614.
[12]谭津,邓长虹,杨威,等. 微电网光伏发电的Adaboost天气聚类超短期预测方法[J]. 电力系统自动化,2017,41(21):33-39.
[13]WANG Y Y,SUN S F,CHEN X Q,et al. Short-term load forecasting of industrial customers based on SVMD and Xgboost[J]. International journal of electrical power and energy systems,2021,129(1):106830.
[14]CHENG R Y,YU J Q,ZHANG M. Short-term hybrid forecasting model of ice storage air-conditioning based on improved SVR[J]. Journal of building engineering,2022,50:104194.
[15]顾海艳,曹林,朱涛,等. 神经网络二次集成预测算法研究[J]. 南京师大学报(自然科学版),2022,45(2):136-141.
[16]ZHANG Q,WU J J,MA Y,et al. Short-term load forecasting method with variational mode decomposition and stacking model fusion[J]. Sustainable energy grids and networks,2022,30:100622.
[17]LIU Y H,NA X,YIN C C,et al. 3-D Joint inversion of airborne electromagnetic and magnetic data based on local Pearson correlation constraints[J]. IEEE transactions on geoscience and remote sensing,2022,60(7):5910813.

备注/Memo

备注/Memo:
收稿日期:2022-09-27.
基金项目:国家自然科学基金项目(61971228).
通讯作者:项晓宇,工程师,研究方向:新能源. E-mail:418892683@qq.com
更新日期/Last Update: 2023-06-15