julius-br/GARFAB

--- license: mit --- **Published**: September 21th, 2022 <br> **Author**: Julius Breiholz # GARFAB-Dataset The (G)erman corpus of annotated (A)pp (R)eviews to detect (F)eature requests (A)nd (B)ug reports (GARFAB) is a dataset to fine-tune models regarding classification of app reviews (ASRs) into "Feature Requests", "Bug Reports" and "Irrelevants" for the German language. All ASRs were collected from the Google Play Store and were classified manually by two independent annotators. A weighted and a full version are published with the following distributions of ASRs: | | Feature Request | Bug Reports | Irrelevant | Total | | --- | --- | --- | --- | --- | full | 345 | 387 | 2212 | 2944 | weighted | 345 | 345 | 345 | 1035 |

原始信息汇总

GARFAB-Dataset 概述

数据集基本信息

• 发布日期: 2022年9月21日
• 作者: Julius Breiholz
• 许可证: MIT

数据集描述

• 目的: 用于微调模型，以将德语应用评论（ASRs）分类为“功能请求”、“错误报告”和“无关内容”。
• 数据来源: 所有ASRs均收集自Google Play Store。
• 分类方式: 由两名独立注释者手动分类。

数据集版本与分布

• 全量版本:

  • 功能请求: 345条
  • 错误报告: 387条
  • 无关内容: 2212条
  • 总计: 2944条

• 加权版本:

  • 功能请求: 345条
  • 错误报告: 345条
  • 无关内容: 345条
  • 总计: 1035条