论文信息 - Convergence Failures in Logistic Regression

Convergence Failures in Logistic Regression

A frequent problem in estimating logistic regression models is a failure of the likelihood maximization algorithm to converge. In most cases, this failure is a consequence of data patterns known as complete or quasi-complete separation. For these patterns, the maximum likelihood estimates simply do not exist. In this paper, I examine how and why complete or quasi-complete separation occur, and the effects they produce in output from SAS ® procedures. I then describe and evaluate several possible solutions.

Paul D. Allison | P. Allison

[1] A. Raftery,et al. Discharge Rates of Medicare Stroke Patients to Skilled Nursing Facilities: Bayesian Logistic Regression with Unobserved Heterogeneity , 1996 .

[2] M. Schemper,et al. The application of Firth's procedure to Cox and logistic regression , 2001 .

[3] A. Albert,et al. On the existence of maximum likelihood estimates in logistic regression models , 1984 .

[4] Thomas J. Santner,et al. A note on A. Albert and J. A. Anderson's conditions for the existence of maximum likelihood estimates in logistic regression models , 1986 .

[5] John S. J. Hsu,et al. Hierarchical Bayesian semiparametric procedures for logistic regression , 1997 .

[6] Anastasios A. Tsiatis,et al. Median Unbiased Estimation for Binary Data , 1989 .

[7] M. Schemper,et al. A solution to the problem of separation in logistic regression , 2002, Statistics in medicine.

[8] Micah Altman,et al. Convergence Problems in Logistic Regression , 2004 .

[9] S. Chib,et al. Bayesian analysis of binary and polychotomous response data , 1993 .

[10] V. Rich. Personal communication , 1989, Nature.

[11] D. Firth. Bias reduction of maximum likelihood estimates , 1993 .

[12] D. Cox,et al. The analysis of binary data , 1971 .