Math Atom · M-D11

Logistic regression

10 minMath AtomStatistics

By the end you will be able to

  1. 1Move between probability, odds and log-odds, and write down the logistic regression model for a 0/1 outcome.
  2. 2Read a coefficient as a change in log-odds and as an odds ratio, and turn a fitted model into predicted probabilities.
  3. 3Explain how the model is fitted by maximum likelihood, and what Newton-Raphson does.
  4. 4Use standard errors and the Wald test to judge whether a characteristic matters.