基于谱聚类的照明行业上市公司业绩分析  

Performance Analysis of Listed Companies in the Lighting Industry Based on Spectral Clustering

作　　者：舒服华 SHU Fuhua(School of Economics of Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]武汉理工大学经济学院,湖北武汉430070

出　　处：《中国照明电器》2025年第1期1-6,16,共7页China Light & Lighting

基　　金：湖北省自然科学基金(2022CFB232)。

摘　　要：客观对照明上市企业经营业绩进行聚类分析,研究其经营业绩的相似性和差异性,对于总结经验,吸取教训,整改完善,不断提升企业的经营水平和盈利能力,促进整个行业高质量发展具有积极的意义。谱聚类是一种建立在谱图理论基础上的聚类算法,与传统的聚类算法相比,它可以在任意形状的样本空间聚类,并将高维空间数据降至低维空间数据处理,具有适应性强、收敛性好、计算简便、易于获得全局最优解等优点。运用谱聚类对我国照明行业部分上市公司的业绩进行了分析,结果表明:分类结果客观准确,内涵深邃丰富,决策过程通俗易懂,生动形象,直观透彻。Objective clustering analysis of the operating performance of listed lighting companies,studying the similarities and differences in their operating performance,is of positive significance for summarizing experience,learning from lessons,rectifying and improving,continuously improving the operational level and profitability of enterprises,and promoting high-quality development of the entire industry.Spectral clustering is a clustering algorithm based on the theory of spectrograms.Compared with traditional clustering algorithms,it can cluster in any shape of sample space and reduce high-dimensional spatial data to low dimensional spatial data processing.It has the advantages of strong adaptability,good convergence,simple calculation,and easy acquisition of global optimal solutions.The performance of some listed companies in China’s lighting industry was analyzed by using spectral clustering.The results indicate that the classification results are objective and accurate,with profound and rich connotations,and the decision-making process is easy to understand,vivid and clear,and intuitive.

关 键 词：照明 上市公司 业绩 分析 谱聚类 

分 类 号：F22[经济管理—国民经济]