To be implemented in Python.

Dataset looks like this in the csv file:

userID	placeID	rating
U1001	H132830	2
U1001	H132825	3
U1001	H135085	1
U1001	H135040	2
U1001	H135039	2
U1001	H135045	2
U1001	H135033	2
U1001	H135025	3
U1001	H135051	2
U1002	H132921	3
U1002	H135062	2
U1002	H135106	2
U1002	H132825	3
U1002	H135052	2
U1002	H132862	3
U1002	H135059	2

Not all entries but you get the idea.

Answer the below recommender systems related questions using the provided ratings dataset (a) For a given user, Ui, estimate the rating for the provided place, Pi, based on the ratings of the k-nearest neighbors who rated that place (i.e., first filter other users based on whether they already rated place P, then identify the k-nearest neighbors from the filtered list). If U; already rated P, then simply return the U;'s rating for Pi. Some tips on the implementation, sample function call and the expected outputs are provided below. # HINT: first generate the pivot table, then calculate distances for knn # de = pd.read_csv ("ratings_data.csv") # df_matrix df.pivot (index = "userID", columns = "placeID", values - "rating ").fillna (0) pred_rating, list_nsimilar_users getPredictedRatings_knn ("ratings_data.csv", user, place, n_neighbors=2) # e.g., for user = "U1001", place "H135080", pred_rating: 1.5, list_nsimilar_users: ['01073', '41057'] = (b) Provide a python function that implements SVD and returns a given number of recommendations along with their estimated ratings for a given user ID. Estimated ratings should NOT be fractional values and can only be one of the following values: 1.0, 2.0, 3.0. That is, you are required to post-process the SVD results to obtain logical estimated ratings (e.g. round to nearest integer). Sample function call and the expected outputs are provided below. dic_place_rating = get RecommendedPlaces_SVD ("ratings_data.csv", "U1077", num_recommendations=3) # dic_place_rating: # {'1132825': 3.0, '1135085': 2.0, '1135038' : 2.0} 2