Logistic Regression
A statistical model used to predict the probability of a binary categorical outcome based on one or more predictor variables.
What is Logistic Regression?
Instead of predicting a continuous value, logistic regression models the log-odds of a categorical dependent variable (usually binary, like pass/fail or yes/no) occurring. It uses a logistic function to squeeze the predicted output to be strictly between 0 and 1, representing a probability.
Why Logistic Regression Matters
It is the standard method for classification problems in research, allowing researchers to estimate how changes in predictors influence the likelihood of a specific event occurring.
Example
A medical researcher uses logistic regression to predict the probability of a patient developing heart disease (yes/no) based on their age, cholesterol levels, and smoking status.
Common Mistakes
- Interpreting the coefficients as linear changes in probability rather than changes in log-odds.
- Using it on highly imbalanced datasets without adjustments, causing the model to simply predict the majority class.