1. the accuracy of a kNN fit on training data (the data used to fit the model) will alwaysdecrease as the k increases. Why or why not?

2. I can only get the classes of points with kNN, never probabilities. (For example, I cant get theprobability of Y=1, only a prediction of either a 0 or 1.) Explain.

3. You and your friend are both running kNN with k=2 on the same data sets and discover you get slightly different accuracy values depending on whether you use your laptop of your friends. However, when you run k=3, you get exactly the same accuracy values with no discrepancies. Should you be concerned about your k=2 results? Why or why not?