Araştırma Makalesi

INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction

Cilt: 9 Sayı: 2 27 Ekim 2026
PDF İndir
TR EN

INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction

Öz

Accurate and generalizable predictive models are essential for modern transport systems, where data are inherently complex, dynamic, and prone to overfitting. This paper proposes Interacted Neural Generalization (INTERGEN), a performance-guided recursive weight-space fusion method that merges independently trained neural networks into a single deployable model. The method iteratively fuses pairs of networks using validation-driven convex parameter combinations, re-evaluates the fused network, and reintroduces it into the fusion pool, enabling interaction at the level of internal representations rather than output-only aggregation. Experiments are conducted on two public transportation prediction tasks: urban-area collision classification using the United Kingdom police-reported road safety dataset, and bike-share demand forecasting using the Washington, District of Columbia. Capital Bikeshare dataset. Results across five random seeds show that the proposed fusion consistently narrows the train–test performance gap, indicating improved generalization. INTERGEN achieves the strongest overall performance on the collision classification task relative to competitive ensemble baselines and achieves the highest explained-variance performance on the demand forecasting task while remaining competitive on error-based metrics. By producing a single consolidated network, the approach reduces deployment and inference complexity compared with multi-model ensembles and supports reproducible evaluation on public datasets.

Anahtar Kelimeler

Kaynakça

  1. Ainsworth, S. K., Hayase, J., & Srinivasa, S. (2022). Git Re-Basin: Merging models modulo permutation symmetries. arXiv:2209.04836.
  2. Breiman, L. (1996). Bagging predictors. Machine Learning, 24(2), 123–140.
  3. Department for Transport. (2023). Road Safety Open Data (STATS19): 2023 collision file and coded-data guide, Great Britain.
  4. Dietterich, T. G. (2000). Ensemble methods in machine learning. In Multiple Classifier Systems (pp. 1–15). Springer. https://doi.org/10.1007/3-540-45014-9_1
  5. Fanaee-T, H., & Gama, J. (2013). Event labeling combining ensemble detectors and background knowledge. Progress in Artificial Intelligence, 2(2–3), 113–127. https://doi.org/10.1007/s13748-013-0040-3
  6. Frankle, J., Dziugaite, G. K., Roy, D. M., & Carbin, M. (2020). Linear mode connectivity and the lottery ticket hypothesis. International Conference on Learning Representations.
  7. Freund, Y., & Schapire, R. E. (1997). A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences, 55(1), 119–139.
  8. Garouani, M., Barhrhouj, A., & Teste, O. (2025). XStacking: An effective and inherently explainable framework for stacked ensemble learning. Information Fusion, 124, 103358. https://doi.org/10.1016/j.inffus.2025.103358

Ayrıntılar

Birincil Dil

İngilizce

Konular

Modelleme ve Simülasyon, Ulaşım ve Trafik

Bölüm

Araştırma Makalesi

Yayımlanma Tarihi

27 Ekim 2026

Gönderilme Tarihi

19 Mayıs 2026

Kabul Tarihi

3 Eylül 2026

Yayımlandığı Sayı

Yıl 2026 Cilt: 9 Sayı: 2

Kaynak Göster

APA
Ekmekçi, M. (2026). INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction. Akıllı Ulaşım Sistemleri ve Uygulamaları Dergisi, 9(2), 370-387. https://doi.org/10.51513/jitsa.1954998
AMA
1.Ekmekçi M. INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction. Jitsa. 2026;9(2):370-387. doi:10.51513/jitsa.1954998
Chicago
Ekmekçi, Mustafa. 2026. “INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction”. Akıllı Ulaşım Sistemleri ve Uygulamaları Dergisi 9 (2): 370-87. https://doi.org/10.51513/jitsa.1954998.
EndNote
Ekmekçi M (01 Ekim 2026) INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction. Akıllı Ulaşım Sistemleri ve Uygulamaları Dergisi 9 2 370–387.
IEEE
[1]M. Ekmekçi, “INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction”, Jitsa, c. 9, sy 2, ss. 370–387, Eki. 2026, doi: 10.51513/jitsa.1954998.
ISNAD
Ekmekçi, Mustafa. “INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction”. Akıllı Ulaşım Sistemleri ve Uygulamaları Dergisi 9/2 (01 Ekim 2026): 370-387. https://doi.org/10.51513/jitsa.1954998.
JAMA
1.Ekmekçi M. INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction. Jitsa. 2026;9:370–387.
MLA
Ekmekçi, Mustafa. “INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction”. Akıllı Ulaşım Sistemleri ve Uygulamaları Dergisi, c. 9, sy 2, Ekim 2026, ss. 370-87, doi:10.51513/jitsa.1954998.
Vancouver
1.Mustafa Ekmekçi. INTERGEN: Performance-Guided Recursive Weight Fusion for Transportation Prediction. Jitsa. 01 Ekim 2026;9(2):370-87. doi:10.51513/jitsa.1954998