ارسل ملاحظاتك

ارسل ملاحظاتك لنا







Sparsity in Bayesian Elastic Net in Tobit Regression with an Application

المصدر: مجلة القادسية للعلوم الإدارية والاقتصادية
الناشر: جامعة القادسية - كلية الادارة والاقتصاد
المؤلف الرئيسي: Flaih, Ahmad Naeem (Author)
مؤلفين آخرين: Alsafi, Mohammed Rasool (Co-Author)
المجلد/العدد: مج23, ع4
محكمة: نعم
الدولة: العراق
التاريخ الميلادي: 2021
الصفحات: 232 - 238
ISSN: 1816-9171
رقم MD: 1235380
نوع المحتوى: بحوث ومقالات
اللغة: الإنجليزية
قواعد المعلومات: EcoLink
مواضيع:
كلمات المؤلف المفتاحية:
Bayesian Estimation | Tobit Regression | Elastic Net Method | Gibbs Sampler Algorithm
رابط المحتوى:
صورة الغلاف QR قانون
حفظ في:
المستخلص: In this paper, we developed one of the most well-known regularization methods that is called elastic net method in tobit regression from the Bayesian point of views. This regularization adding the ridge and lasso penalty functions to the residual sum of squares term. In this paper, we developed new Bayesian hierarchical model for the tobit regression based on the proposed scale mixture of Li and Lin, (2010) that mixing the normal distribution with truncated gamma distribution (1,∞) as double exponential prior distribution for the interested parameter (β). Furthermore, the MCMC Gibbs sampling algorithm has developed for the posterior distribution of interested parameter (β). Analysis of real data has conducted for the proposed model; also, a comparative has made with some regression models. The proposed model is outperformed and gives promised results.

ISSN: 1816-9171

عناصر مشابهة