پیوندهای مفید
آمار
| تعداد نشریات | 32 |
| تعداد شمارهها | 591 |
| تعداد مقالات | 5,779 |
| تعداد مشاهده مقاله | 8,660,782 |
| تعداد دریافت فایل اصل مقاله | 6,342,639 |
Residual Bootstrap Inference for Testing Points of Impact in Functional Linear Regression | ||
| Caspian Journal of Mathematical Sciences | ||
| مقالات آماده انتشار، پذیرفته شده، انتشار آنلاین از تاریخ 14 شهریور 1405 | ||
| نوع مقاله: Research Articles | ||
| شناسه دیجیتال (DOI): 10.22080/cjms.2026.31979.1859 | ||
| نویسنده | ||
| Alireza Shirvani* | ||
| Department of Statistics, Faculty of Basic Sciences, Velayat University, Iranshahr, Iran | ||
| تاریخ دریافت: 11 خرداد 1405، تاریخ بازنگری: 23 مرداد 1405، تاریخ پذیرش: 27 مرداد 1405 | ||
| چکیده | ||
| Functional linear regression models with points of impact provide an important framework for identifying localized effects of functional predictors. Existing inference procedures are mainly based on asymptotic approximations which may exhibit unsatisfactory finite-sample performance. In this paper, a residual bootstrap procedure is developed for testing points of impact in functional linear regression models. The asymptotic validity of the proposed bootstrap calibration is established under general Gaussian functional predictors. It is shown that the bootstrap distribution consistently approximates the limiting null distribution of the test statistic. Simulation studies demonstrate that the proposed bootstrap approach substantially improves empirical significance levels and finite-sample accuracy compared with classical asymptotic approximations, particularly under highly dependent covariance structures. A real data illustration based on the Canadian Weather data set further demonstrates the practical applicability of the proposed bootstrap methodology and confirms its effectiveness in finite-sample inference. | ||
| کلیدواژهها | ||
| Bootstrap inference؛ Functional regression؛ Point of impact models؛ Gaussian processes؛ Functional data analysis | ||
|
آمار تعداد مشاهده مقاله: 4 |
||