Speakers: Henry Qiao ), Ningde New Energy Technology Co., Ltd., 6Sigma/TRIZ Promoter
the speech topic : Apply Firth Penalized maximum likelihood estimation in GLM Solve the problem of data separation in
Summary of speech:
In binomial data modeling, it is often encountered that the value range of p is [0, 1], and the value range of the general linear model response 𝑦 is (−∞,+∞); the relationship between p and factor x is not Linear; p is for the case where the value variance of different factors x is not equal. When there are many explanatory variables, the sample size is small, and the data distribution is unbalanced, the fitting is prone to separation problems. Applying Firth penalty maximum likelihood estimation is a better way to solve the data separation problem. The basic idea of Firth Penalized Maximum Likelihood Estimation (PMLE=Penalized Maximum Likelihood Estimation) is to add a penalty term to the logarithmic likelihood function, thereby reducing the deviation of the parameter maximum likelihood estimate. Maximum Likelihood Estimation MLE Based on the asymptotic theory, it depends on the layout of each data point in the sample space; when the sample size is large and the data distribution is balanced, better results can be obtained. This video will detail how to apply Firth penalized maximum likelihood estimation to solve the data separation problem in GLM through practical examples.
