用于端到端无监督决策树的核 KMeans 聚类分裂
Kernel KMeans clustering splits for end-to-end unsupervised decision trees
AI总结：
针对无标签数据的端到端聚类树学习挑战，提出无监督二叉树模型 Kauri，通过贪婪最大化核 KMeans 目标且无需定义质心，在多数据集上表现优异。
摘要：
树模型是在相对较小的数据集上获得可解释预测的便捷模型。尽管在监督学习中有许多关于端到端构建此类树的提议，但在没有标签的情况下进行端到端聚类学习树模型仍是一个开放性挑战。由于大多数工作侧重于用树来解释另一种聚类算法的结果，我们在此提出一种新颖的端到端训练的无监督二叉聚类树：Kauri。该方法执行核 KMeans 目标的贪婪最大化，而无需定义质心。我们在多个数据集上将此模型与近期的无监督树进行比较，并表明当使用线性核时，Kauri 表现相同。对于其他核，Kauri 通常优于核 KMeans 和 CART 决策树的拼接。
英文摘要：
Trees are convenient models for obtaining explainable predictions on relatively small datasets. Although there are many proposals for the end-to-end construction of such trees in supervised learning, learning a tree end-to-end for clustering without labels remains an open challenge. As most works focus on interpreting with trees the result of another clustering algorithm, we present here a novel end-to-end trained unsupervised binary tree for clustering: Kauri. This method performs a greedy maximisation of the kernel KMeans objective without requiring the definition of centroids. We compare this model on multiple datasets with recent unsupervised trees and show that Kauri performs identically when using a linear kernel. For other kernels, Kauri often outperforms the concatenation of kernel KMeans and a CART decision tree.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
