DETERMINACIÓN DE LA CONSTANTE K EN LA REGRESIÓN RIDGE
ABSTRACTSince the Ordinary Least Square (OLS), no have inside its structure an optimisation method for determine the effect that the multicollinearity has over the estimate coefficients of the vector , permit to the Ridge Regression (RR) take an important roll in solve this problem. In this paper, w...
Đã lưu trong:
| Những tác giả chính: | , , |
|---|---|
| Định dạng: | Artículo |
| Ngôn ngữ: | spa |
| Được phát hành: |
Universidad Autónoma de Ciudad Juárez
2015
|
| Những chủ đề: | |
| Truy cập trực tuyến: | http://erevistas.uacj.mx/ojs/index.php/culcyt/article/view/538 |
| Các nhãn: |
Thêm thẻ
Không có thẻ, Là người đầu tiên thẻ bản ghi này!
|
| Tóm tắt: | ABSTRACTSince the Ordinary Least Square (OLS), no have inside its structure an optimisation method for determine the effect that the multicollinearity has over the estimate coefficients of the vector , permit to the Ridge Regression (RR) take an important roll in solve this problem. In this paper, we present the steps for the determination of the constant of proportionality K, with we obtain a smaller variance that this estimate by OLS, over the focus of the Total mean square of the Prediction (TMSP). β ˆ RESUMEN Desde que el método de Mínimos Cuadrados (MC), no tiene dentro de su estructura un método de optimización para determinar el efecto que la multicolinealidad tiene sobre los coeficientes estimados del vector , permite a la regresión Ridge (RR), tomar importancia en resolver este problema. En este artículo, presentamos el desarrollo de la determinación de la constante de proporcionalidad K, con la que obtenemos una varianza más pequeña que la estimada por MC, bajo el enfoque del Cuadrado Medio de la Predicción Total (CMPT). β ˆPalabras Claves: Cuadrado Medio de la Predicción, Multicolinealidad, Regresión Ridge, Mínimos Cuadrados, |
|---|---|
| số ISSN: | 2007-0411 |