Again another problem with terminology and Machine Learning.
Since Logistic Regression is a classification model dealing here with a categorical target class, Why isn’t it called Logistic Classification?
Yeah this is an unfortunate historical thing … if you find a good reference giving more details don’t hesitate to add a comment about in this topic.
1 Like
Maybe because logistic regression estimates the parameters of a statistical model to model a binary dependent variable.
For example, when using a logistic function for the statistical model, one assumes a linear relationship between the predictor variables and the logarithm of the odds.
That is the logistic model: y = log_base(p / (1-p)) = b_0 + b_1 * x_1+ … + b_n * x_n.
y is a Bernoulli variable (yes/no, 1/0, class A/not class A) taking the value 1 with probability p and 0 with probability (1 - p).
The regression coefficients, b_i’s, have to be estimated.
The best explanation I’ve found is in Quora by Mr Raymond Anderson, Why isn't Logistic regression called Logistic Classification? - Quora