在非寿险分类费率厘定中,存在各种模型可供选择,如加法模型、乘法模型、混合模型和广义线性模型等,而在这些模型的参数估计中,还存在各种可供选择的估计方法,如最小二乘法、极大似然法、最小x2法、直接法和边际总和法等。这些模型和参数估计方法散见于各种精算学文献中,本文对这些模型和参数估计方法进行了系统的比较和分析,并揭示了它们之间存在的一些等价关系。
In non - life insurance, various models are selectable to calculate relativities, such as additive models, multiplicative models, mixed models and GLM. For estimating the parameters of these models, different methods may be used, such as minimum least squares, maximum likelihood, minimum x^2, straight method and marginal total. The paper systematically analyzes these models and estimation methods and reveals some of their relationships.