Binary Logistic Regression

Binary logistic regression is a generalized linear model that predicts the probability of a two-category outcome from one or more predictors by modeling the log-odds as a linear function.

이항 로지스틱 회귀분석은 로그오즈를 선형함수로 모형화하여 하나 이상의 예측변수로 두 범주 결과의 확률을 예측하는 일반화선형모형이다.

How do I model which factors predict a yes/no outcome?

예/아니오 결과를 예측하는 요인을 어떻게 모형화하나요?

Binary logistic regression estimates how predictors change the log-odds of a binary outcome such as success versus failure. It fits any dichotomous dependent variable, and its exponentiated coefficients are reported as odds ratios.

When to use Binary Logistic Regression

• A dichotomous outcome such as disease/no disease or purchase/no purchase
• Estimating adjusted odds ratios for several predictors
• Predicting class probabilities for a binary classification problem

When to use something else

• The outcome has three or more unordered categories: use multinomial logistic regression
• The outcome has ordered categories: use ordinal logistic regression

Assumptions and requirements

• Independent observations
• Linearity between continuous predictors and the log-odds
• No severe multicollinearity
• Sufficient events per predictor to avoid separation and overfitting

What the analysis reports

• Coefficients (log-odds) and odds ratios with confidence intervals
• Wald or likelihood-ratio tests
• Pseudo R² (e.g., Nagelkerke)
• Hosmer–Lemeshow test
• Classification table or ROC curve

How to report it

[Predictor] was significantly associated with higher odds of [outcome], OR = [value], 95% CI [[lower], [upper]], p = [value].

Related methods

• Multinomial Logistic Regression
• Probit Regression
• ROC-AUC Curve Analysis

Run Binary Logistic Regression on Quorant