How to Classify Decision Boundaries

Linear classifiers learn linear decision boundaries and so that's what's happening is those boundaries are being defined. Changing the intercepts change the y intercept of the boundary. So think of the boundary as a slope line where you're either moving the y intercept up and down the y axis or you're changing the rise over run or slope. You can then plot those the decision boundary using plot underscore classifier and then you can look at the error.