Logistic Regression

Despite its name, Logistic Regression is a method for classification. It is conceptually similar to linear regression. However, because we want to model discrete categories, the logistic function (a sigmoid curve) represents alternating between categories better.

While building the model, a threshold value where the sidmoid curve sharply raises is optimized.

In R, the generalized linear model function glm() creates logistic regressions:

model <- glm(Y ~.,family=binomial(link='logit'), data=training)

summary(model)

result <- predict(model,newdata=Xtest,type='response')

result <- ifelse(result > 0.5,1,0)

Hints

• One key advantage of the method is that you obtain coefficients for each input feature that are easy to interpret.
• Each coefficient is accompanied by a probability that describes statistical significance
• Logistic Regression can handle multiple categories
• also see https://www.r-bloggers.com/how-to-perform-a-logistic-regression-in-r/