Seed-Guided Semi-Supervised Clustering by A-Contrario Anomaly Detection
Researchers have developed a novel semi-supervised clustering framework that leverages the statistical duality between grouping principles and anomaly detection. This method, called "clustering-by-exclusion," uses a Perception algorithm with an expectation-based threshold to identify outliers without manual parameter tuning. By employing minimal user-provided seeds, the algorithm iteratively refines clusters, effectively isolating noise and identifying new clusters, and has shown competitive performance on various benchmarks. AI
IMPACT This new clustering approach could improve data analysis and pattern recognition in machine learning tasks.