Seeded Hierarchical Clustering for Expert-Crafted Taxonomies

TL;DR

This paper introduces HierSeed, a weakly supervised hierarchical clustering algorithm that efficiently assigns unlabeled documents to expert-crafted taxonomies using minimal labeled seeds, outperforming existing methods.

Contribution

HierSeed is a novel, data-efficient, and computationally lightweight algorithm for seed-based hierarchical clustering of unlabeled data.

Findings

HierSeed outperforms unsupervised baselines on three datasets.

HierSeed surpasses supervised methods in accuracy.

The approach is both scalable and effective with limited labeled data.

Abstract

Practitioners from many disciplines (e.g., political science) use expert-crafted taxonomies to make sense of large, unlabeled corpora. In this work, we study Seeded Hierarchical Clustering (SHC): the task of automatically fitting unlabeled data to such taxonomies using only a small set of labeled examples. We propose HierSeed, a novel weakly supervised algorithm for this task that uses only a small set of labeled seed examples. It is both data and computationally efficient. HierSeed assigns documents to topics by weighing document density against topic hierarchical structure. It outperforms both unsupervised and supervised baselines for the SHC task on three real-world datasets.