Get Started
Home
Topics
Search
Library
Research questionHow should limited samples be allocated across classes to estimate class-weighted risk for rare events in dependent sequences?Rare positive cases can dominate a class-weighted loss, while dependence among observations makes sampling variability and train-test separation harder to control. The allocation must therefore balance information from both strata without assuming independent observations.
Economics
Finance
Machine Learning
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.Optimal Stratified Allocation for Rare-Event Onset Forecasting in Dependent SequencesThe evidence concerns sampling K much smaller than a finite labelled population, with class-conditional sampling without replacement and a weighted risk estimator. The application forecasts statistically explosive price-regime onsets in 350 U.S. equities from 2004–2011, with fewer than 1% positive rows and five purged forward blocks; the reported allocation ordering matches predictions at a 10-day horizon, while its predicted dependence on prevalence across horizons does not.research paper · Sep 3, 2026
Related questions
How should imbalanced multi-label data be sampled to support rare-label inference while accounting for target frequencies and label dependencies?How can quantile risk control meet predefined risk thresholds with finite-sample guarantees without excessive conservatism?How can deep imbalanced regression avoid underfitting scarce tail labels when uncertainty varies across instances?How can extreme quantile treatment effects be estimated under heavy tails while respecting location invariance?