1. GLM stands for "generalized linear model" and extends linear regression beyond models with normal error distributions. Logistic regression and other exponential family distributions are examples of GLMs.
2. Logistic regression models the logit (log odds) as a linear function of predictors using the logit link function, allowing probabilities to be estimated.
3. The term "linear" refers to the model being linear in its parameters, not necessarily that the regression line is straight. Non-linear terms like quadratic and cubic terms can be included while keeping the model linear.