摘要
在以往关于模型选择的研究中,一般都是先假定选定一个模型,然后对由此模型确定的分布族进行比较,求出最优的分布函数和数值特征,忽略了模型本身的不确定性.介绍了Bayes方法在模型选择中的方法及应用,举例说明了用Bayes方法选择模型,不仅能够减小模型选择中模型不确定性的影响,而且可以根据实际情况和问题认识程度的深化,对模型进行扩展.
In many papers about model selection, a common approach is that a model is selected firstly and so a kind of distributions is determined.Then the best distribution is chosen from them. But the uncertainty of model is ignored.This paper presents the theory and application of the Bayesian method in model selection. It proved that Bayesian method can diminish the influence of the uncertainty of model, and can also expand model easily on the basis of the actual use and further insight into the question.
出处
《郑州大学学报（工学版）》
CAS
2003年第2期93-95,共3页
Journal of Zhengzhou University（Engineering Science）
