报告人：梁成安副教授 澳门大学
报告题目：Probabilistic interpretations of Bayesian Information Criteria (BIC)
时 间：2016-06-07 (星期二) 10:10 ~ 11:40
地 点：旗山校区理工北楼601报告厅
主办：数学与计算机科学学院, 福建省分析数学及应用重点实验室, 数学研究中心
参加对象：统计、数学教育方向的老师和学生
报告摘要：Bayesian Information Criteria (BIC) is usually used by many researchers ONLY as an indicator in model selection.However, Bayesianly, i.e. from the Bayesian perspective, BIC can serve as an approximation towards the posterior probabilities, and the differences in BIC can serve as the log of the Bayes Factors, which gives a much more meaningful uses and interpretations.We shall illustrate the uses of BIC through a small scale simulation studies, which demonstrate a thorough analysis of the inter-relationship of 3 variables: one for each of the independent, dependent and mediating variables.
专家简介：梁成安，澳门大学副教授、博士生导师，持续教育中心主任，1990年获伦敦经济政治学院统计学博士学位。主要研究教育测量与评价、潜变量模型、模型拟合度、统计在社会科学研究中的应用，在Computational Statistics & Data Analysis、British Journal of Mathematical and Statistical Psychology、Journal of Family Nursing、The Journal of Experimental Education、Journal of Social Service Research、Psychologia等国际期刊上发表多篇学术论文。
