In this problem, we will take a different approach towards feature selection and choosing the best value of . We will be working with the california housing dataset. Recall that in class we used the correlation between median prices and features to pick the most predictive features. Using correlation, we selected the following features: MedInc HouseAge AveRooms Latitude Longitude Now we will do a more refined feature selection by using the 2 of knn models with different features. We will fit 3 different model families: knn with varying k, and features = 'Latitude' and 'Longitude' knn with varying k, and features = 'Latitude', 'Longitude', 'MedInc' knn with varying k, and features = 'Latitude', 'Longitude', 'MedInc' and 'HouseA