DeepAI AI Chat
Log In Sign Up

Impossibility of Depth Reduction in Explainable Clustering

by   Chengyuan Deng, et al.
Rutgers University

Over the last few years Explainable Clustering has gathered a lot of attention. Dasgupta et al. [ICML'20] initiated the study of explainable k-means and k-median clustering problems where the explanation is captured by a threshold decision tree which partitions the space at each node using axis parallel hyperplanes. Recently, Laber et al. [Pattern Recognition'23] made a case to consider the depth of the decision tree as an additional complexity measure of interest. In this work, we prove that even when the input points are in the Euclidean plane, then any depth reduction in the explanation incurs unbounded loss in the k-means and k-median cost. Formally, we show that there exists a data set X in the Euclidean plane, for which there is a decision tree of depth k-1 whose k-means/k-median cost matches the optimal clustering cost of X, but every decision tree of depth less than k-1 has unbounded cost w.r.t. the optimal cost of clustering. We extend our results to the k-center objective as well, albeit with weaker guarantees.


page 7

page 9

page 13


How to Find a Good Explanation for Clustering?

k-means and k-median clustering are powerful unsupervised machine learni...

Explainable k-means. Don't be greedy, plant bigger trees!

We provide a new bi-criteria Õ(log^2 k) competitive algorithm for explai...

Near-optimal Algorithms for Explainable k-Medians and k-Means

We consider the problem of explainable k-medians and k-means introduced ...

The computational complexity of some explainable clustering problems

We study the computational complexity of some explainable clustering pro...

ExKMC: Expanding Explainable k-Means Clustering

Despite the popularity of explainable AI, there is limited work on effec...

Near-Optimal Explainable k-Means for All Dimensions

Many clustering algorithms are guided by certain cost functions such as ...

Non-Depth-First Search against Independent Distributions on an AND-OR Tree

Suzuki and Niida (Ann. Pure. Appl. Logic, 2015) showed the following res...