Key points about logistic regression:

• It is used when the dependent variable is categorical (binary or multiclass classification) rather than continuous.

• It predicts the probability of an observation belonging to a specific class rather than the class itself.

• The relationship between independent variables X and dependent variable Y is modelled using a logistic/sigmoid curve rather than linear function.

• The output is a probability value between 0 and 1, via the logistic function formula:

P(Y=1|X) = 1/(1+e^-(b0 + b1X1 + b2X2 +…))

• The coefficients b have similar interpretations as linear regression – effect of one unit change in X on log odds of Y.

• It can handle both continuous and categorical independents.

• Can be extended for multiclass classification problems using techniques like one-vs-rest.

• Evaluated using classification metrics like accuracy, AUC-ROC, confusion matrix etc. rather than R-squared.

• Commonly used for problems like prediction, recommendation systems, sentiment analysis etc.

So in summary, logistic regression applies when we want to model and predict categorical outcomes using linear regression techniques for multiple variables.