Model

The Stepwise Model Builder uses the linear logistic model.

This model has a dichotomous dependent variable, and in this module the outcome is labeled as either having a Good or Bad outcome. Included in the Stepwise Model Builder dialog are options for Dependent Variable (Y), Good code and Bad code. For the model, the dependent variable is assumed to have a probability π, where for the ith case:

For n observations, y₁ through y_n, with probabilities π_i through π_n and case weights w₁ through w_n, the likelihood function is:

The logarithm of l is:

The derivative of L with respect to ß_j is:

Did you find this helpful?

Yes No