查詢結果分析
來源資料
相關文獻
- 貝氏廣義隱藏式類別分析法在心理計量學的應用
- The Comparison of BIC and Corrected BIC Criteria for Model Selection of Time Series
- Finite Mixture Multivariate Generalized Linear Models Using Gibbs Sampling and E-M Algorithms
- 長期資料迴歸均數函數的決定
- An Asymptotic Theory for Linear Model Selection
- Information Criteria for Multiple Data Sets and Restricted Parameters
- Bootstrap Estimate of Kullback-Leibler Information for Model Selection
- A Unified Approach to Estimating and Modeling Linear and Nonlinear Time Series
- A Bootstrap Variant of AIC for State-Space Model Selection
- 廣義階層線性模式在強制汽車責任保險理賠資料之應用
頁籤選單縮合
題 名 | 貝氏廣義隱藏式類別分析法在心理計量學的應用=The Application of Bayesian Extended Latent Class Analyses in Psychometrics |
---|---|
作 者 | 楊志堅; 楊志強; | 書刊名 | 測驗年刊 |
卷 期 | 46:2 1999.07[民88.07] |
頁 次 | 頁73-84 |
分類號 | 319.711 |
關鍵詞 | 吉氏取樣; 貝氏廣義隱藏類別分析; 模式選取; 酒精上癮; Gibbs sampling; Bayesian extended latent class analysis; Model selection; Alcohol dependence; |
語 文 | 中文(Chinese) |
中文摘要 | 吉氏取樣(Gibbs sampling)是近幾年來極受矚目的統計計算法之一,它讓貝氏統計(Bayesian Statistics)及推論可以在有限的計算資源(computing resources)下被實際運用的,更重要的是它的多樣性(flexibility),使許多以前幾乎不可能付諸實用的統計模式成為可能,這使得它成為近幾年研究貝氏統計的最主要工具。利用吉氏取樣來做貝氏隱藏類別分析(Bayesian Latent Class Analysis, BLCA)及貝氏廣義隱藏類別分析(Bayesian Extended Latent Class Analysis, BELCA)不僅在統計計算學上有其重要性而且在醫學統計或心理計量學上亦有其實用價值。本文以酒癮關聯模式之瞼例研究來說明如何以吉氏取樣來實踐BLCA及BELCA模式,並討論其估算過程(estimation procedure)、收斂診斷(convergence diagnosis)及模式選取(model selection)等主題研究結果顯示BLCA及BELCA確是一可供心理計量學上實際推廣運用的有效統計模式。 |
英文摘要 | Theis paper introduces latent class analysis and extended latent class analysis models by using Gibbs sampling to show the significance of the two models in practical areas of statistical computation, biomedical statistics and psychometrics. The Gibbs sampling method for Bayesian implementation reasons in recent years. One of the major advancements by using Gibbs sampling is that Gibbs sampling allows researchers to execute many complex Bayesian statistical models practically because of its flexibility. This important feature also makes Gibbs sampling become one of the most popular tools for Bayesian modeling. We further demonstrate the procedures of parameter estimation, convergence diagnosis, and model selection of both models engaging Gibbs sampling by using a practical example of NLSY alcohol dependence study. |
本系統中英文摘要資訊取自各篇刊載內容。