7. (14 points) Machine Learning Potpourri: Answer the questions below (a) (3 points) Given the figure below match the regions with the "fit type. 50 Training Performance Testing Performance 40 B Error % A 10 0 2 8 10 6 Parameter Good Fit: Under Fit: Over Fit: (b) (2 points) Given only the above figure, what value (approximately) would you chose for the parameter? (c) (3 points) You want to use linear regression to fit a function to the data given below where a is the input and y is the output. How would you formulate this, i.e., what would be the form of the function you fit? Make sure to clearly mark the parameters and inputs. (d) (2 points) You are given the below data points (21, 22) and their corresponding classes, marked by their color (blue vs orange) and marker shape (dot vs X). Is the kNN method feasible for solving this classification problem using only the given inputs (11, 12)? 3 2 N 1 t 0 -1 -2 -3 -3 -2 -1 o 1 3 3 N X1 (e) (2 points) Come up with a feature that would be useful for this classification problem? (e.g. assume you want to solve it with a linear method) a