多维属性融合构建亲和度的谱聚类分析算法研究
打开文本图片集
中图分类号：TP311 文献标识码：A文章编号：2096-4706（2025）21-0122-05
Research on Spectral Clustering Analysis Algorithm of Constructing Affinity by Fusing Multi-dimensional Attributes
YANG Jingting，WANGXiangyu，MAHongdan，LIANGRuiwen，GE Dongxu （NanjingAuditUniversity Jinshen College，Nanjing21oo23，China）
Abstract： Inthecontext ofthedeep integrationof information technologies and social networks，multi-dimensional data hasshownexplosivegrowth，anditscharacteristicsofighdimensionalitycomplexstructureandidencorelationvebecoe thecore problemsinthefieldofdataanalysis.Inthefaceofsuch problems，traditionalclustering methodsoftenhavepoor clusteringefectuetooblemsschsmensioaldudcyigncantsarsitndiglesiilarityeauretich arediffcult tomeet theneedsofactualsenarios.hisstdyistoextractseveralindependentfactorswithhighcoelationand weak corelationbetweenfactorsbasedonfactoranalysis methodtoreducethedimensionalityofmulti-dimensionaldata，select apropriate affnityalgorithms fordiffrenttypesofdata，constructacompositeaffitymatrix with dynamicweightsand fusion ofmultipleafnitylgorithms，ndgenerateananitymatrixsuitableforspecificapplicationsenarios，whichisfinallused forspectralclusteringanalysis.This methodcannotonlyimprove theclusteringaccuracy，butalsoprovide moreauratedata support for application scenarios such as personalized recommendation and public opinion analysis.
Keywords： factor analysis; spectral clustering; affnity; multi-dimensional data; dynamic weight
0 引言
随着信息技术的飞速发展和网络信息的不断丰富，复杂的多维数据信息越来越庞杂，如何有效处理和分析这些数据成为信息管理与信息系统领域的一大挑战。（剩余6720字）
