Understanding how a logistic regression classifier works
Although this section name sounds a bit unheard, it is correct. Logistic regression is indeed a regression model, but it is mostly used for classification tasks. A classifier is a model that contains sets of rules or formulas (sometimes millions or more) to perform the classification task. In a simple logistic regression classifier, we only need one rule built on a single feature to perform the classification.
Logistic regression is very popular in both traditional statistics as well as machine learning.
The name logistic originates from the name of the function used in logistic regression: logistic function. Logistic regression is the Generalized Linear Model (GLM). The GLM is not a single model, but an extended group of models of Ordinary Least Squares (OLS) models. Roughly speaking, the linear part of the model in GLM is similar to OLS, but various kinds of transformation and interpretations are introduced, so GLM models...
