Cometa: Multi-Label Classification Data Archive for Machine Learning

Cometa is a repository of multi-label datasets created by fdavidcl and last updated in March 2019. It provides a centralized archive for machine learning researchers focusing on multi-label classification problems.

Use Cases

• Benchmarking algorithms against multi-label targets
• Analyzing label dependencies within multi-label classification tasks
• Training supervised models on datasets with multiple output categories

Strengths

• Specialized focus on multi-label classification
• Open-source availability under AGPL-3.0
• Curated specifically for machine learning benchmarking

Limitations

• Significant lack of metadata regarding dataset scale and features
• Temporal staleness with no updates since 2019
• Unknown data quality or noise levels due to missing documentation

Provenance

Source

fdavidcl via GitHub

Freshness

Last updated in March 2019; no recent maintenance or additions are documented.