Decision boundaries

In the following, you will use Paint Data to paint data, Predictions to see how well the model fit the data (do not mind about overfitting here), Scatter Plot to observe decision boundaries and Tree Viewer to observe the tree.

Use the Paint Widget to draw some data the is

• easy to model using logistic regression and difficult for trees (tree may still be able to achieve high accuracy because we're fitting and testing on the same data, but observing the tree in Tree View will show its struggle);
• easy for trees and difficult for logistic regression (you can observe the decision boundary in the scatter plot, and the simplicity of the Tree in Tree Viewer);
• data that is difficult for both (here, you should be able to make the logistic regression's results terrible, while the tree's will still be OK, but the tree is going to be large).