.png){kind=icon}
Return
Machine Learning Classification and Portfolio Construction: Does the Loss Function Matter?
.png){kind=icon}
PRE
AI
DOI：10.1080/0015198x.2026.2726133.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Classification outperforms regression across matched machine learning models in portfolio construction. A stacking ensemble of gradient boosted trees, random forest, and neural network yields a value-weighted annualized Sharpe ratio of 2.08 for classification and 1.39 for regression. This outperformance strengthens with class granularity and persists across subsamples and after transaction costs. Spanning tests show that classification retains economically large alphas after we control for regression, whereas regression alphas shrink substantially once we control for classification. These results indicate that classification extracts more return information than matched regression. Our diagnostics trace classification’s advantage to more precise separation of return deciles.
Keywords:
asset pricing
classification
controlled experiment
loss function
machine learning
portfolio construction
regression
2.0
Journal
F
IF:
2.2
Papers:
1.2K
Citations:
3.1K
Organization
No organization information available
Cited Papers
No cited papers available