基于Lasso和K-means聚类的宏观经济周期阶段划分  

Macroeconomic Cycle Stage Division Based on Lasso and K-means Clustering

作　　者：李绮雯 赵国瑞 田婷婷 曾烁飒 邱福权 LI Qi-wen;ZHAO Guo-rui;TIAN Ting-ting;ZENG Shuo-sa;QIU Fu-quan(School of Business,Guangdong Ocean University,Yangjiang 529599,Guangdong,China;School of Computer Science and Engineering,Guangdong Ocean University,Yangjiang 529599,Guangdong,China;School of Materials Science and Engineering,Guangdong Ocean University,Yangjiang 529599,Guangdong,China)

机构地区：[1]广东海洋大学商学院,广东阳江529599 [2]广东海洋大学计算机科学与工程学院,广东阳江529599 [3]广东海洋大学材料科学与工程学院,广东阳江529599

出　　处：《喀什大学学报》2024年第6期30-34,共5页Journal of Kashi University

基　　金：广东省哲学社会科学“十三五”规划2020年度学科共建项目“基于机器学习算法的高校‘学困生’精准识别与预警研究”(GD20XJY20);广东省大学生创新创业训练计划项目“教育大数据视阈下高校“学困生”精准识别与学习预警模型研究”(S202310566090).

摘　　要：经济周期划分对于把握市场经济规律本质具有重要意义.对经济周期划分的研究,国内外主要从经济学理论出发,对其质性进行研究.但影响经济周期划分的潜在因素繁多,且影响机制复杂,因此目前对经济周期划分的定量研究不足,尤其在经济学理论与机器学习等理论与方法的交叉研究方面有待加强.为此,从近20年的宏观经济数据出发,对高维宏观经济指标加入惩罚Lasso进行筛选,遴选出采购经理指数、居民消费指数等关键指标,继而进行K-means聚类,以对我国2001—2021年经济周期进行划分.The division of economic cycles is of great significance for grasping the essence of the laws of market economy.In this field,domestic and foreign research mainly focuses on qualitative research based on economic theory.However,there are many potential factors affecting the division of economic cycles,and the influencing mechanism is complex,so there are few quantitative studies,especially the interdisciplinary research with machine learning and other theories and methods needs to be strengthened.Based on the macroeconomic data of the past 20 years,this paper selects the high-dimensional macroeconomic indicators to add the penalty Lasso to the Key indicators,and then selects the purchasing managers'index and household consumption index as the Key indicators,and then performs K-means clustering to obtain the division of China’s economic cycle from 2001 to 2021.

关 键 词：经济周期划分 机器学习 Lasso K-MEANS聚类 

分 类 号：O213[理学—概率论与数理统计] F124.8[理学—数学]