当前位置: X-MOL 学术Adv. Data Anal. Classif. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Functional data clustering by projection into latent generalized hyperbolic subspaces
Advances in Data Analysis and Classification ( IF 1.6 ) Pub Date : 2021-01-07 , DOI: 10.1007/s11634-020-00432-5
Alex Sharp , Ryan Browne

We introduce a latent subpace model which facilitates model-based clustering of functional data. Flexible clustering is attained by imposing jointly generalized hyperbolic distributions on projections of basis expansion coefficients into group specific subspaces. The model acquires parsimony by assuming these subspaces are of relatively low dimension. Parameter estimation is done through a multicycle ECM algorithm. Application to simulated and real datasets illustrate competitive clustering capabilities, and demonstrate the models general applicability.



中文翻译:

通过投影到潜在广义双曲子空间中的功能数据聚类

我们介绍了一个潜在的子空间模型,该模型有助于基于模型的功能数据聚类。通过将基本扩展系数的投影上的联合广义双曲分布施加到组特定子空间中,可以实现灵活的聚类。该模型通过假设这些子空间的维数相对较小来获得简约性。参数估计是通过多周期ECM算法完成的。在模拟和真实数据集上的应用说明了竞争性聚类能力,并证明了模型的一般适用性。

更新日期:2021-01-07
down
wechat
bug