I have a business application where I want to make fast classifications, and update my model quickly and immediately when a new training case comes along. Which of the following algorithms is best suited?

K$-$nearest neighbor

Naive Bayes classifier

Linear regression

Logistic regression