20-MAD (20-MAD: Mozilla Apache Dataset)

20-MAD，一个连接 Mozilla 和 Apache 项目的提交和发布数据的数据集。它包含超过 20 年的 765 个项目、340 万次提交、230 万次问题和 1730 万次问题评论的信息，其压缩大小超过 6 GB。数据包含有关源代码提交（例如，添加和删除的行、消息和提交时间）和问题（状态、严重性、投票和摘要）的所有典型信息。问题评论已针对自然语言处理和情感分析进行了预处理。这包括表情符号以及效价和唤醒分数。

20-MAD is a dataset that compiles commit and release data from both Mozilla and Apache projects. It contains information spanning over 20 years across 765 projects, including 3.4 million commits, 2.3 million issues, and 17.3 million issue comments, with a compressed size exceeding 6 GB. The dataset includes all typical information related to source code commits, such as lines added and deleted, commit messages and commit timestamps, as well as issue-related details including status, severity, vote counts, and summaries. Issue comments have been preprocessed for natural language processing (NLP) and sentiment analysis, with preprocessing covering emoji handling as well as valence and arousal scores.

搜集汇总

背景与挑战

背景概述

20-MAD是一个涵盖Mozilla和Apache项目超过20年开发历史的数据集，包含765个项目、340万次提交、230万次问题及1730万次问题评论，数据量超过6GB。它提供了源代码提交和问题的详细信息，其中问题评论经过预处理，适用于自然语言处理和情感分析任务。