Chp 9: Logistic Regression

What is the primary purpose of logistic regression?

To model binary target variables and predict probabilities.

What does logistic regression use instead of ordinary least squares?

Maximum Likelihood Estimation (MLE).

What type of curve does logistic regression produce?

An S-shaped (sigmoid) curve.

Why is linear regression not suitable for binary targets?

It can produce predicted values outside the range of 0 to 1.

What is heteroskedasticity in the context of linear regression?

The assumption that error terms have constant variance, which is violated in binary outcomes.

What is the range of predicted values in logistic regression?

Between 0 and 1.

What is the logit function?

The log odds of the probability, used in logistic regression.

What does the inverse logit function do?

Transposes the logit function to estimate probabilities.

How does a 1 unit increase in SAT score affect the odds in logistic regression?

It increases the odds by a factor represented by the odds ratio.

What is an odds ratio?

A measure of how the odds of an event occurring change with a one unit increase in a variable.

What is the significance of the cutoff value in logistic regression?

It determines the threshold for classifying predicted probabilities into binary outcomes.

What are the two main tasks logistic regression can be used for?

Explanatory tasks (profiling) and predictive tasks (classification).

What does the term 'non-conforming probabilities' refer to?

Predicted probabilities that fall outside the valid range of 0 to 1 in linear regression.

What is the relationship between predictors and the response variable in logistic regression?

They are related via a nonlinear function called the logit.

What is the main issue with using linear regression for binary outcomes?

It leads to poor approximations of the relationship between data values.