multinomial logistic regression
Multinomial logistic regression is a statistical classification method for predicting one outcome from three or more categories that are treated as mutually exclusive. It extends logistic regression by estimating how input features relate to the probabilities of the available classes, typically using one class as a reference for expressing the other classes' scores. The model converts those scores into class probabilities, which can support a predicted category as well as an estimate of uncertainty. Its coefficients can offer insight into how features are associated with class outcomes, subject to the model's assumptions and chosen reference category. It is a useful baseline for multi-class tasks, particularly when interpretability matters, though its performance may be limited when the underlying decision boundaries are highly complex.