A solution to the problem of separation in logistic regression.
نویسندگان
چکیده
The phenomenon of separation or monotone likelihood is observed in the fitting process of a logistic model if the likelihood converges while at least one parameter estimate diverges to +/- infinity. Separation primarily occurs in small samples with several unbalanced and highly predictive risk factors. A procedure by Firth originally developed to reduce the bias of maximum likelihood estimates is shown to provide an ideal solution to separation. It produces finite parameter estimates by means of penalized maximum likelihood estimation. Corresponding Wald tests and confidence intervals are available but it is shown that penalized likelihood ratio tests and profile penalized likelihood confidence intervals are often preferable. The clear advantage of the procedure over previous options of analysis is impressively demonstrated by the statistical analysis of two cancer studies.
منابع مشابه
Using Classic Discriminant Analysis and Detection Function for Separation of Chemical Victims in Sardasht City to Exposed and Non Exposure Mustard Groups in 2013 and Comparison with Logistic Regression
متن کامل
Sample size determination for logistic regression
The problem of sample size estimation is important in medical applications, especially in cases of expensive measurements of immune biomarkers. This paper describes the problem of logistic regression analysis with the sample size determination algorithms, namely the methods of univariate statistics, logistics regression, cross-validation and Bayesian inference. The authors, treating the regr...
متن کاملSimulation of Store Separation using Low-cost CFD with Dynamic Meshing
The simulation of the store separation using the automatic coupling of dynamic equations with flow aerodynamics is addressed. The precision and cost (calculation time) were considered as comparators. The method used in the present research decreased the calculation cost while limiting the solution error within a specific and tolerable interval. The methods applied to model the aerodynamic force...
متن کاملThe Effect of Transitive Closure on the Calibration of Logistic Regression for Entity Resolution
This paper describes a series of experiments in using logistic regression machine learning as a method for entity resolution. From these experiments the authors concluded that when a supervised ML algorithm is trained to classify a pair of entity references as linked or not linked pair, the evaluation of the model’s performance should take into account the transitive closure of its pairwise lin...
متن کاملA hybrid metaheuristic algorithm for solving a rollon–rolloff waste collection vehicle routing problem considering waste separation and recycling center
This paper addresses a rollon–rolloff waste collection vehicle routing problem which involves large containers that collect huge amounts of waste from construction sites and shopping malls. In this problem, vehicles could only move one container at a time between customers’ locations, a depot, a disposal center, and a recycling center site. Other than a disposal center, a recycling ...
متن کاملNon-Fourier heat conduction equation in a sphere; comparison of variational method and inverse Laplace transformation with exact solution
Small scale thermal devices, such as micro heater, have led researchers to consider more accurate models of heat in thermal systems. Moreover, biological applications of heat transfer such as simulation of temperature field in laser surgery is another pathway which urges us to re-examine thermal systems with modern ones. Non-Fourier heat transfer overcomes some shortcomings of Fourier heat tran...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Statistics in medicine
دوره 21 16 شماره
صفحات -
تاریخ انتشار 2002