LogisticRegressionModel Class

Use the LogisticRegressionModel class to analyze a situation where the outcome can have two or more possible values. A logistic regression model tries to express one variable, called the dependent variable, which can have only two distinct values, as a function of one or more other variables called independent variables or predictors in a specific form.

Logistic regression is a special case of a GeneralizedLinearModel with a binomial distribution and the logit link function. To perform variants of logistic regression, like probit regression, use the GeneralizedLinearModel class.

In addition to binary logistic regression, the LogisticRegressionModel can also represent multinomial logistic regression, where there may be more than two outcomes. In this case, the dependent variable must be a CategoricalVariable.