{"version":1,"type":"story","url":"https://digestai.news/story/researchers-release-160-000-article-brazilian-portuguese-financial-new","json":"https://digestai.news/story/researchers-release-160-000-article-brazilian-portuguese-financial-new.json","markdown":"https://digestai.news/story/researchers-release-160-000-article-brazilian-portuguese-financial-new.md","slug":"researchers-release-160-000-article-brazilian-portuguese-financial-new","headline":"Researchers release 160,000-article Brazilian Portuguese financial news dataset","summary":"Researchers have introduced a new large-scale corpus containing approximately 160,000 financial news articles in Brazilian Portuguese. This dataset aims to address the scarcity of domain-specific data for Financial Natural Language Processing (NLP) in the region. The study focuses on establishing initial benchmarks for weakly supervised multi-class classification using semantic similarity.\n\nTo evaluate performance, the team annotated subsets of the corpus both manually and using a Large Language Model (LLM). They tested three similarity-based methods: cosine similarity, K-Nearest Neighbors (KNN), and FAISS-based nearest-neighbor search, utilizing multilingual semantic embeddings. The experiments highlighted that the classification task is highly sensitive to label quality and class distribution.\n\nResults showed that KNN achieved the best performance on the manually labeled subset, with 47.3% accuracy and a 31.9% F1-score. In contrast, while accuracy reached approximately 50% on the LLM-annotated subset, F1-scores dropped to near 20%. This discrepancy underscores the negative impact of label imbalance and noisy weak annotations, emphasizing the critical need for high-quality manual annotations in future research.","keyPoints":["New dataset contains approximately 160,000 Brazilian Portuguese financial news articles.","KNN achieved 47.3% accuracy on manually labeled data, outperforming other methods.","LLM-annotated data showed high accuracy but F1-scores dropped to near 20%."],"whyItMatters":"This dataset provides a foundational resource for developing NLP systems in Brazilian Portuguese, a language with limited financial data. It highlights the limitations of weak supervision and the necessity for high-quality manual annotations in financial classification tasks.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":[],"models":[],"people":[]},"firstPublishedAt":"2026-10-16T04:00:00Z","updatedAt":"2026-10-16T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"arXiv cs.CL","title":"A Large-Scale Dataset for Multi-Class Financial News Classification in Brazilian Portuguese","url":"https://arxiv.org/abs/2610.20630","publishedAt":"2026-10-16T04:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Researchers release 160,000-article Brazilian Portuguese financial news dataset\", 16 October 2026, https://digestai.news/story/researchers-release-160-000-article-brazilian-portuguese-financial-new","publisher":"Digest AI","title":"Researchers release 160,000-article Brazilian Portuguese financial news dataset","datePublished":"2026-10-16T04:00:00Z","url":"https://digestai.news/story/researchers-release-160-000-article-brazilian-portuguese-financial-new"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}