TOPICS
Search

Conditional Logit Regression


Conditional logit regression assumes a model of the form

 p_j=(e^(beta^'x_j))/(sum_(j)e^(beta^'x_j))

for j=1, ..., k+1. In this model, a subject is presented with choice alternatives and asked to choose the best alternative. In addition, the set of alternatives is typically the same for all subjects and the explanatory variables are all choice-specific.


Explore with Wolfram|Alpha

Cite this as:

Weisstein, Eric W. "Conditional Logit Regression." From MathWorld--A Wolfram Web Resource. https://mathworld.wolfram.com/ConditionalLogitRegression.html

Subject classifications