A new R package, discoCVI, has been released to efficiently evaluate density-based clusterings, including the quality of noise labels. This package implements the DISCO index, which uses a single-linkage merge tree to reduce memory requirements from quadratic to linear with the number of points, while maintaining a quadratic running time. Benchmarks show discoCVI is significantly faster and more memory-light than previous versions and its Python counterpart, capable of scoring 100,000 points in approximately 50 seconds with minimal memory usage. AI
IMPACT Provides a more efficient tool for researchers working with density-based clustering algorithms.
RANK_REASON The cluster describes a new software package and its implementation for a specific research task (clustering evaluation), published on arXiv. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →