Systematic empirical program beyond one-dimensional illustrations

Design a systematic empirical program to evaluate Fibonacci ensembles on high-dimensional regression datasets, binary and multiclass classification with margin-based losses, structured data modalities including images, sequences, and graphs, and rigorous comparisons against stacking, Super Learner, and deep ensemble baselines, while preserving the conceptual clarity of the Fibonacci architecture.

Background

The paper presents controlled one-dimensional regression experiments with random Fourier features and polynomial bases to illustrate bias–variance and spectral effects of Fibonacci weighting.

The Future Work section calls for a broader empirical program across high-dimensional, classification, and structured-data scenarios, along with comprehensive comparisons to modern ensemble baselines, to assess performance and robustness in practical settings.

References

Several directions remain open and are, in our view, both challenging and promising.

— On Fibonacci Ensembles: An Alternative Approach to Ensemble Learning Inspired by the Timeless Architecture of the Golden Ratio  (2512.22284 - Fokoué, 25 Dec 2025) in Section “Future Work: From One-Dimensional Harmony to High Dimensional Practice”