Gibbs Sampler方法在考生写作能力的贝叶斯估计中的应用

Bayesian Inference in Examinees'Writing Ability Using the Gibbs Sampler

本文将同族ＴＥＳＴ模型（Ｊｏｒｅｓｋｏｇ，１９７１）一般化之后作为作文考试数据的模型，并应用ＧｉｂｂｓＳａｍｐｌｅｒ方法对考生的作文能力进行贝叶斯估计。通过具体实例分析，说明了这种方法的估计是合理的，得到的结果是有效的。

In this paper, we consider an extented congeneric test model for essay tests.Themarginal postior distribution of examinees'writing ability can be esimated by using the GibbsSampler.The application of the method to the real data resulted in reasonable estmates.