Economy Watchers Survey
收藏资源简介:
经济观察者调查数据集由东京大学和北海道大学的研究团队构建,利用日本中央政府机构发布的材料。该数据集包含超过300,000个样本,涉及当前和未来经济评估的文本和标签数据。数据集的创建过程包括自动更新框架,确保最新数据集的可用性。该数据集主要应用于日本金融领域的自然语言处理任务,如分类和情感分析,旨在解决金融领域中语言模型的评估问题。
The Economic Observer Survey Dataset was developed by a research team from The University of Tokyo and Hokkaido University, utilizing materials released by Japanese central government agencies. This dataset contains over 300,000 samples, including text and labeled data related to current and future economic assessments. An automatic update framework is incorporated into the dataset's creation process to ensure the availability of the latest dataset. Primarily applied to natural language processing tasks in the Japanese financial domain, such as classification and sentiment analysis, this dataset is designed to address the problem of language model evaluation in the financial sector.
数据集概述
基本信息
- 语言: 日语
- 许可证: CC-BY 4.0
- 数据规模: 100K<n<1M
数据配置
- 配置名称: current
- 训练集:
current/train.jsonl - 验证集:
current/validation.jsonl - 测试集:
current/test.jsonl
- 训练集:
- 配置名称: future
- 训练集:
future/train.jsonl - 验证集:
future/validation.jsonl - 测试集:
future/test.jsonl
- 训练集:
使用方法
python
需要 datasets >= 2.15.0
from datasets import load_dataset ds = load_dataset( "retarfi/economy-watchers-survey", name="current", revision="2024.06.0", split="validation", )
- name: 可选项为
"current"(当前经济周期) 或"future"(未来经济周期)。 - revision: 如未指定,则读取最新数据。
- split: 如指定,数据以
datasets.Dataset格式读取,否则以datasets.DatasetDict格式读取。
许可证
CC-BY 4.0

- 1Economy Watchers Survey provides Datasets and Tasks for Japanese Financial Domain东京大学 · 2024年



