基于广义Frechet距离的区间值函数型聚类方法OACHSSCD

Interval-valued functional clustering is a statistical analysis method that reveals the intrinsic structure of inter-val-valued functional data.The existing interval-valued functional clustering methods typically use absolute distances between function curves as similarity measures,neglecting the shape features and structural information of the function curves.These meth-ods are often influenced by data dimensionality and outliers,resulting in suboptimal clustering outcomes.In order to address the above deficiencies,this paper proposes a novel interval-valued functional clustering method.The method is based on the general-ized Frechet distance to measure the similarity between function curves and expresses distance information in interval form,which better captures the trend of function curve variations.Additionally,a tournament algorithm is introduced to enhance clustering effi-ciency.In the empirical studies,clustering analysis is conducted on temperature data from Chinese cities by using the proposed method,and the results are compared with clustering outcomes based on functional Manhattan distance and interval-valued func-tional Euclidean distance.The empirical results indicate that the proposed method outperforms other approaches in interval-val-ued functional clustering tasks.