This tutorial covers how to find the decision boundary in logistic regression classification with two features.

  1. Select data series 'Class1' and data series 'Class2' on left panel [Points].
  2. Select menu item <Analysis><LogisticRegression>.
  3. Training dataset will be divided on two parts based on parameter Z value.
  4. In dialog on step 1 it will assign Z1=0 to points from Class1 and Z2=1 to points from Class2.
  5. We find the decision boundary curve line in form of X,Y polynomials.
  6. On step 2 select checked 'Reularization' and set parameter Lambda value to 20.
  7. Vary polynomial power parameters M and N to find best fit but to keep simple formula.
  8. to minimize percent of incorrect classified items in training and testing datasets.
  9. On step 2 select Class1 color, points in Class2 wll be brushed with inverted color.
  10. Close the dialog.