logit_model: the individualized binary logistic regression for categorical response
data.
Description
logit_model fit the categorical data by the individualized binary
logistic regression
Usage
logit_model(splitted, newY)
Arguments
splitted
A list containing the datasets which we will use in the
categorical case. Note that the element of the splitted is the collections
of samples from Classes 0 and Classes k.
newY
A numeric number denotes the value of the labels from 0 to K
which is the number of categories
Value
beta_mat
a matrix contains the estimated coefficient. Note that the
beta_mat is a n * p matrix which n is the number of the explanatory variables
and p+1 is the number of categories
Details
logit_model fits the splitted data by using the the individualized binary
logistic regression according to the value of newY. Because we use use Class
0 as the baseline for modeling the probability ratio of Class k to Class 0 by
fitting K individual logistic models, if newY equal to 0, it means we need
fit all elements of the splitted data. Otherwise, we only fit the samples
from class 0 and class newY.
References
Li, J., Chen, Z., Wang, Z., & Chang, Y. I. (2020). Active learning in
multiple-class classification problems via individualized binary models.
Computational Statistics & Data Analysis, 145, 106911.
doi:10.1016/j.csda.2020.106911