A Large-Scale Dataset for Multi-Class Financial News Classification in Brazilian Portuguese

查看原文 / View Original 编辑 / Edit

Computer Science > Computation and Language arXiv:2610.20630 (cs) [Submitted on 29 Sep 2026] Title: A Large-Scale Dataset for Multi-Class Financial News Classification in Brazilian Portuguese Authors: Leomax Bandeira, Gabriel Aguiar, Dalton Cézane Gomes Valadares, Luiz Pereira, Helder Fernando de Araújo Oliveira, Gustavo Xavier, Andson M. Balieiro.

计算机科学 > 计算与语言 arXiv:2610.20630 (cs) [提交于 2026 年 9 月 29 日] 标题：用于巴西葡萄牙语多类金融新闻分类的大规模数据集作者：Leomax Bandeira, Gabriel Aguiar, Dalton Cézane Gomes Valadares, Luiz Pereira, Helder Fernando de Araújo Oliveira, Gustavo Xavier, Andson M. Balieiro。

Abstract: The development of Financial Natural Language Processing (NLP) systems for Brazilian Portuguese is hindered by the scarcity of domain-specific datasets. To address this, we present a large-scale corpus of approximately 160,000 Brazilian Portuguese financial news articles collected from web sources.

摘要：针对巴西葡萄牙语的金融自然语言处理（NLP）系统的开发受到领域特定数据集匮乏的阻碍。为了解决这一问题，我们提出了一个包含约 16 万篇从网络来源收集的巴西葡萄牙语金融新闻文章的大规模语料库。

To establish initial benchmarks for a weakly supervised multi-class classification approach based on semantic similarity, we annotated subsets of this corpus both manually and using a Large Language Model (LLM). We evaluated three similarity-based methods — cosine similarity, K-Nearest Neighbors (KNN), and FAISS-based nearest-neighbor search — using multilingual semantic embeddings.

为了建立基于语义相似度的弱监督多类分类方法的初步基准，我们通过人工和大型语言模型（LLM）对该语料库的子集进行了标注。我们使用多语言语义嵌入评估了三种基于相似度的方法——余弦相似度、K-近邻（KNN）和基于 FAISS 的最近邻搜索。

Our experiments reveal that the task is highly sensitive to label quality and class distribution. On the manually labeled subset, KNN achieved the best performance (47.3% accuracy, 31.9% F1-score), closely followed by FAISS. Conversely, on the LLM-annotated subset, while accuracy reached approximately 50%, F1-scores dropped to near 20%, highlighting the negative impact of label imbalance and noisy weak annotations.

我们的实验表明，该任务对标签质量和类别分布高度敏感。在人工标注的子集上，KNN 取得了最佳性能（47.3% 的准确率，31.9% 的 F1 分数），FAISS 紧随其后。相反，在 LLM 标注的子集上，虽然准确率达到了约 50%，但 F1 分数降至接近 20%，这凸显了标签不平衡和嘈杂的弱标注所带来的负面影响。

These findings establish foundational baselines for a challenging task, underscoring the critical need for high-quality manual annotations. Ultimately, this dataset and our preliminary results provide a reference point for future research in Brazilian Portuguese Financial NLP.

这些发现为一项具有挑战性的任务建立了基础基准，强调了高质量人工标注的迫切需求。最终，该数据集和我们的初步结果为巴西葡萄牙语金融 NLP 的未来研究提供了参考点。