Design-based individual prediction
This work addresses the need for reliable error estimation in predictive modeling, particularly in contexts with complex sampling, but it appears incremental as it builds on existing design-based inference methods.
The paper tackles the problem of making valid inference for individual prediction errors by developing a design-based approach that accounts for sampling and cross-validation designs, treating outcomes and features as constants.
A design-based individual prediction approach is developed based on the expected cross-validation results, given the sampling design and the sample-splitting design for cross-validation. Whether the predictor is selected from an ensemble of models or a weighted average of them, valid inference of the unobserved prediction errors is defined and obtained with respect to the sampling design, while outcomes and features are treated as constants.