L3Cube-MahaNews: News-based Short Text and Long Document Classification Datasets in Marathi
Saloni Mittal, Vidula Magdum, Omkar Dhekane, Sharayu Hiwarkhedkar, Raviraj Joshi
TL;DR
This work addresses the scarcity of large-scale, multi-length Marathi text classification data by introducing L3Cube-MahaNews, a 12-label corpus with over 1.08 lakh records across short headlines, long paragraphs, and long documents derived from Lokmat. It provides rigorous benchmarks using three BERT variants (MahaBERT, IndicBERT, MuRIL) and demonstrates that a monolingual Marathi model (MahaBERT) generally yields superior performance across all dataset lengths, with long documents (LDC) offering the strongest fine-tuning signal. The authors thoroughly describe dataset curation, data collection, statistics, and evaluation, revealing that datasets exhibit length-dependent performance and limited cross-dataset transfer, underscoring the need for length-diverse, multi-dataset training. The resources are publicly released to support Marathi NLP research and practical topic classification applications, with acknowledged limitations related to OCR content and paragraph-level label noise for future refinement.
Abstract
The availability of text or topic classification datasets in the low-resource Marathi language is limited, typically consisting of fewer than 4 target labels, with some achieving nearly perfect accuracy. In this work, we introduce L3Cube-MahaNews, a Marathi text classification corpus that focuses on News headlines and articles. This corpus stands out as the largest supervised Marathi Corpus, containing over 1.05L records classified into a diverse range of 12 categories. To accommodate different document lengths, MahaNews comprises three supervised datasets specifically designed for short text, long documents, and medium paragraphs. The consistent labeling across these datasets facilitates document length-based analysis. We provide detailed data statistics and baseline results on these datasets using state-of-the-art pre-trained BERT models. We conduct a comparative analysis between monolingual and multilingual BERT models, including MahaBERT, IndicBERT, and MuRIL. The monolingual MahaBERT model outperforms all others on every dataset. These resources also serve as Marathi topic classification datasets or models and are publicly available at https://github.com/l3cube-pune/MarathiNLP .