Researchers have introduced a new problem in machine learning called Extreme Binary Classification, focusing on classifiers with an extremely low false negative rate. To tackle this, they developed a threshold adaptation method grounded in Extreme Value Theory and a feature selection technique using permutation tests on sample maxima. Experiments on four datasets showed the approach outperforms current methods and demonstrated its utility in cancer screening. AI
IMPACT Introduces a novel approach to classification problems with strict false negative constraints, potentially improving accuracy in sensitive applications like medical screening.
RANK_REASON Research paper introducing a new problem and methodology in machine learning. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- cancer screening
- Extreme Binary Classification
- Extreme Value Theory
- machine learning
- permutation test
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
