| dc.contributor | 統計學系 | en_US |
| dc.creator (作者) | 陳麗霞 | zh_TW |
| dc.date (日期) | 1997 | en_US |
| dc.date.accessioned | 2-Sep-2014 17:32:02 (UTC+8) | - |
| dc.date.available | 2-Sep-2014 17:32:02 (UTC+8) | - |
| dc.date.issued (上傳時間) | 2-Sep-2014 17:32:02 (UTC+8) | - |
| dc.identifier.uri (URI) | http://nccur.lib.nccu.edu.tw/handle/140.119/69618 | - |
| dc.description.abstract (摘要) | 本研究旨在探討迴歸模式之解釋變數存在測量誤差時,如何配置解釋變數的值,以使得在某一準則之下,下一個反應變數與其目標值的差異為最小。我們首先討論簡單線性迴歸架構下,最小平方確定等價(LSCE)控制法則的漸近特性。接著以貝氏架構闡釋如何以Gibbs抽樣法估計下一個反應變數的預測期望損失(PEL),並且選取使PEL為最小的解釋變數值為貝氏控制的配置值。當貝氏控制的解析形式不存在時,則以蒙第卡羅最小化法估計之,並討論貝氏控制估計氏的漸近性質。< | en_US |
| dc.description.abstract (摘要) | In this research, we discuss how to select settings of a regressor when it is subject to measurement errors, in order to minimize the difference between the next response and its target under certain criterion. We first discuss the asymptotic behaviors of the least squares certainty equivalence (LSCE) control rule. Then based on the Bayesian framework, we illustrate how to estimate the predictive expected loss (PEL) via Gibbs sampling for the next response. The Bayes control rule is the one minimize the PEL. When the analytical form of the Bayes rule is not available, then Monte Carlo minimization is employed to find a minimizer of the estimated PEL, and the asymptotic properties of such minimizer is studied.< | en_US |
| dc.format.extent | 490 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.language.iso | en_US | - |
| dc.relation (關聯) | 行政院國家科學委員會 | en_US |
| dc.relation (關聯) | 計畫編號NSC86-2115-M004-007 | en_US |
| dc.subject (關鍵詞) | 最小平方確定均等;線性迴歸;測量誤差;預測期望損失;Gibbs抽樣;蒙地卡羅最小化 | en_US |
| dc.title (題名) | 線性迴歸模式具測量誤差下控制問題之研究 | zh_TW |
| dc.title.alternative (其他題名) | Control Problem for Linear Regression Models with Measurement Errors. | en_US |
| dc.type (資料類型) | report | en |
