中文摘要：

非寿险分类费率的厘定通常采用的方法有单项分析法、最小偏差法和广义线性模型，特别是后面两种方法在非寿险实务中应用十分广泛，精算文献中对这两种方法的理论和应用研究也较多，但对二者的比较研究较少。本文首先对最小偏差模型和广义线性模型进行了简要介绍，之后对这两种分类费率模型进行了系统的比较研究，总结了它们各自的优缺点以及二者之间的一些等价关系，最后通过一组实际的汽车保险数据讨论了它们的应用。

英文摘要：

In non-life insurance, one-way analysis, are usually used to calculate relativities. In particular, minimum bias models and generalized linear models are widely applied in the practice. The research about them can be found in various actuarial literatures. However, tess ~gtel~ti~m are paid to the comparisol~s of the t~＇o methods. This paper system- atically compares their respective advantages and disadvantages and also summarizes some equivalence relations between the two kinds of models. And finally the paper applies the models to a set el data from auto insurance.