Logistic Regression

A model for categorical response variables with two levels

Logistic Regression

A model for response varaibles where regular multiple regression is ineffective

Generalised Linear Model (GLM)

A transfomation to relate a continuous probability to the predictors in a logistic regression - ln(pi/(1-pi))

Logit Transformation

Presentation of logistic regression results similar to that of multiple regression

Summary Table

Statistic used for variable selection in logistic regression, analogous in application to adjusted r-squared

Akaike Information Criterion (AIC)

Include independence between outcomes and linearity of predictors with logit(pi)

Conditions for Logistic Regression

Checking model quality through diagnostic plots and residual analysis

Model Assessment

Grouping data based on predicted probabilities to aid assessment

Bucketing Data

Graphical representation of residuals for model evaluation

Residual Plot