Random Feature Embeddings Give Strong Baselines for Extreme Multi-label Text Classification

Authors: Asha Rani, Rudra Dutt & Yashaswi Verma

Venue: International Journal of Data Science and Analytics, 21 (1), 27 (2026) — Q2 Journal

We show that random feature embeddings provide strong, computationally efficient baselines for extreme multi-label text classification, outperforming several complex methods on benchmark datasets.