EUVIMEDEUVIMED entdecken
Crossref

Efficient estimation for deep accelerated failure time model with application to credit risk analysis

Kun Ren, Li Liu, Mu He, Wen Su

Electronic Journal of Statistics · 01.01.2026 · 20 (2) · 5580-5617

DOI: 10.1214/26-ejs2605 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronic Journal of Statistics, erschienen am 01.01.2026. Verfasst von Kun Ren und 3 weiteren Autor:innen. Die Arbeit nennt 52 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Time-to-event data, frequently encountered in credit risk analysis and engineering reliability studies, present significant analytical challenges due to censoring and data heterogeneity. We develop a deep semiparametric accelerated failure time model designed for right-censored data, which accommodates complex data structures while maintaining interpretability through parametric components. Our approach integrates the flexibility of deep neural networks with traditional smoothing techniques, preserving the benefits of parametric interpretability. Furthermore, we establish both nonasymptotic and asymptotic properties for the proposed method, including prediction error bounds, asymptotic normality, and semiparametric efficiency of the estimator. Simulation studies demonstrate the superior performance of our method compared to conventional smoothing approaches. Finally, we apply our method to analyze the German credit data.

Zitieren

APA

Ren, K., Liu, L., He, M., & Su, W. (2026). Efficient estimation for deep accelerated failure time model with application to credit risk analysis. Electronic Journal of Statistics, 20(2), 5580-5617. https://doi.org/10.1214/26-ejs2605

Vancouver

Ren K, Liu L, He M, Su W. Efficient estimation for deep accelerated failure time model with application to credit risk analysis. Electronic Journal of Statistics. 2026;20(2):5580-5617. doi:10.1214/26-ejs2605

BibTeX
@article{Ren2026,
  author = {Ren, Kun and Liu, Li and He, Mu and Su, Wen},
  title = {Efficient estimation for deep accelerated failure time model with application to credit risk analysis},
  journal = {Electronic Journal of Statistics},
  year = {2026},
  volume = {20},
  number = {2},
  pages = {5580--5617},
  doi = {10.1214/26-ejs2605}
}
Kostenlos starten →