WASABI Song Corpus

WASABI Song Corpus是由蔚蓝海岸大学创建的大型歌曲数据集，包含173万首带有歌词的歌曲。数据集通过网络音乐数据库提取元数据，并结合歌词和音频分析进行丰富。该数据集涵盖36种不同语言的歌曲，主要用于支持音乐搜索引擎和音乐专业人士处理大量歌词集合，实现智能浏览、分类和歌曲推荐。数据集创建过程中，采用了多种自然语言处理方法和信息提取技术，旨在解决音乐领域中歌词内容的结构化、情感分析和主题识别等问题。

The WASABI Song Corpus is a large-scale song dataset developed by Université Côte d'Azur, comprising 1.73 million songs with lyrics. The dataset extracts metadata from online music databases and is enriched via lyric and audio analysis. It covers songs in 36 distinct languages, and is primarily designed to support music search engines and music professionals in processing large-scale lyric corpora, enabling intelligent browsing, classification, and song recommendation. During its development, the dataset adopts a variety of natural language processing methods and information extraction technologies, aiming to address challenges including lyric content structuring, sentiment analysis, and topic recognition in the music domain.