A general mixture loss function to optimize a personalized predictive model
Tatiana Krikella, Joel A Dubin · 2026
Read A general mixture loss function to optimize a personalized predictive model by Tatiana Krikella, Joel A Dubin (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13688343.txt.
Description
Precision medicine builds a prediction model for one patient from a crowd of similar patients, but how large should the crowd be? Tatiana Krikella and Joel Dubin propose a loss function that tunes that choice directly. The user picks which virtues count, discrimination, calibration, or a blend, and the method searches crowd sizes for the best joint score.
Simulations showed the useful search range runs from twenty to seventy percent of the training data. The optimal size moved with the chosen virtues, proving the setting is not one-size-fits-all. Earlier findings about crowd size and accuracy held up under the new method. The authors aim for personal prediction that is both flexible and cheap to compute.
Catalog record
| CREATOR | Tatiana Krikella, Joel A Dubin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |