Logistic regression model
Modelo de regresión logística
- Date
- 1944 · confidence high
- Place
- Mayo Clinic, Rochester, Minnesota, United States
- Credited to
- Joseph Berkson
- Field
- Knowledge & symbols
What was discovered
Berkson adapted the logistic function, known since the 19th century, to model the probability of a binary outcome from explanatory variables, coining the term 'logit'. From around 1970 it became the standard model for binary regression in medicine, social science and engineering.
The logistic function itself had been introduced by Verhulst in the 19th century to model population growth; Berkson's 1944 contribution was applying it to binary-outcome regression and giving it its modern statistical apparatus.
What it led to
- The standard for binary regression in medicine and social science
- A basis for classification in supervised machine learning
Built on
- Pierre François Verhulst's logistic function (1838-1847)