E. Boj del Val, T. Costa Cor
We estimate the mean squared error of prediction for logistic regression. First, we approximate the mean squared error by the sum of the process variance and of the estimation variance. Then, we obtain an expression for the estimation variance by using the delta method. Finally, prediction error is the square root of the mean squared error of prediction. In previous works we obtained, for generalized linear models, the general formulas of prediction error in the cases of the power family of error distributions and of the power family of link functions. Now, we show the expression of the mean squared error of prediction for the generalized linear model with Binomial error distribution and logit link function, the named logistic regression. We illustrate the interpretability and the usefulness of prediction error when we use logistic regression. We make an example with real data.
Palabras clave: Logistic regression, mean squared error, estimation variance, delta method
Programado
L06.1 Grupo de Análisis Multivariante y Clasificación II
5 de septiembre de 2016 12:55
0.02 - Aula de proyectos 1
