New CARSANN method enhances nearest neighbor classification accuracy

Researchers have developed a new method called Curvature-Aware Radius Shrinkage for Adaptive Nearest Neighbor Classification (CARSANN) to improve the accuracy of nearest neighbor classification. This framework adapts the spatial extent of each neighborhood based on the local geometric complexity of the data, unlike traditional k-NN which uses a fixed neighborhood size. CARSANN estimates intrinsic dimensionality and local mean curvature to control neighborhood scale, shrinking radii in highly curved regions and expanding them in flatter areas. Experiments on OpenML datasets demonstrated that CARSANN significantly boosts balanced accuracy, improving it from 0.6506 to 0.7528 on average across datasets. AI

IMPACT This new method could improve the performance of various machine learning applications that rely on nearest neighbor algorithms.

RANK_REASON The cluster contains a research paper detailing a new classification method. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

• Alexandre Levada
• arXiv
• Friedman test
• k-nearest neighbors algorithm
• Nemenyi test
• OpenML
• TwoNN

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →