方法对比

	奖惩系统	可信度理论	负二项回归
领域	精算学	精算学	计量经济学
方法族	Regression model	Regression model	Regression model
起源年份	1995	1967	2011
提出者	Jean Lemaire	Hans Bühlmann	Hilbe (textbook treatment); generalized linear model framework
类型	Actuarial experience-rating model	Weighted linear blend of individual and collective experience	Generalized linear model for count data
开创性文献	Lemaire, J. (1995). Bonus-Malus Systems in Automobile Insurance. Kluwer Academic Publishers. ISBN: 978-0-7923-9545-5	Bühlmann, H. (1967). Experience rating and credibility. ASTIN Bulletin, 4(3), 199–207.	Hilbe, J. M. (2011). Negative Binomial Regression (2nd ed.). Cambridge University Press.
别名	No-Claim Discount System, Merit Rating System, Experience Rating in Automobile Insurance, Prim-Ceza Sistemi	Bühlmann Credibility, Experience Rating, Linear Credibility Estimator, Güvenilirlik Teorisi	NB regression, NB2 regression, negatif binom regresyonu
相关	2	3	4
摘要	A Bonus-Malus System (BMS) is an actuarial experience-rating mechanism used primarily in automobile insurance to adjust individual policyholders' premiums based on their personal claim history. Policyholders who remain claim-free receive premium discounts (bonus), while those who file claims are penalised with surcharges (malus). The framework was comprehensively formalised and analysed by Jean Lemaire in his landmark 1995 monograph, which remains the definitive reference for the design and evaluation of such systems worldwide.	Credibility Theory is an actuarial framework for estimating the pure premium of an individual risk by blending its own observed loss experience with the collective (portfolio) mean. Introduced by Hans Bühlmann in 1967, the method derives the optimal linear combination—the credibility-weighted premium—that minimises mean squared error. It extends classical experience rating to a rigorous statistical footing rooted in Bayesian and linear estimation principles.	Negative Binomial Regression is a generalized linear model for count outcomes that extends Poisson regression to handle overdispersion, where the variance of the counts exceeds their mean. Developed in the GLM tradition and treated in depth by Hilbe (2011), it adds a dispersion parameter so that inference stays valid when Poisson would understate the spread of the data.