'127': 1,
'56': 5,
'7': 5,
'98': 3,
'237': 5,
'117': 4,
'172': 4,
'222': 5,
'204': 3
}
#comes from Movie_Data_Frames module
rated_movies = ('50', '258', '100', '181', '294', '286', '288', '1', '300',
'121', '174', '127', '56', '7', '98', '237', '117', '172',
'222', '204')
df = build_dataframe(user_dict)
def build_evaluation_file(user_dict, k):
"""Takes dict input of user ratings and returns csv of to calculate error"""
try:
new_df = pd.DataFrame(list(user_dict.items()),
columns=['Movie ID', 'Ratings'])
user_tr = new_df.set_index(
['Movie ID']) #Application user's movie ratings in data frame
#Create transposed versions of these dataframes in order to use the correlation method available in Pandas
#df is the queried dataframe built using the build_dataframe function in line 35
mean_df = df.fillna(
df.mean()
for film in range(len(movie_list)):
if (film < (10 * multiple) and cont.lower()) == 'y':
print(final_picks[film])
elif len(final_picks) > 10:
cont = input("Enter 'y' if you would like to see more films:\n ")
if cont.lower() == 'y':
multiple +=1
else:
return(print("Thank you for using this service."))
break
while True:
user_dict = build_profile() #Function prompts uses to rate 20 movies. The films presented have the largest number of reviews in the database
k = eval(input("How many neighbors should be used for the recommendations? (k)")) #This is used for User CF and Item CF
print("This should take a few seconds...finding great movies for you!")
movie_df = build_dataframe(user_dict) #The dataframe is based on a query in the database getting ratings for all users who rated the same movies as the application's user
cf_neighbors = build_similarity(user_dict, movie_df, k)#returns the nearest neighbors for the app user and their correlations
cf_movies = find_movies(user_dict,cf_neighbors) #returns movies with predicted ratings for the user
final_picks = select_genres(cf_movies)#gives the app's user the choice to request films of a certain genre
print(display_films(final_picks))
e = input("I hoped those movies look good. I'm going to try another way to find movies for you, press enter to start the process:\n")
print("This should take a few seconds...finding great movies for you!")
item_movies = item_item(user_dict, k) #Returns movies that are similar to items rated by the app user along with predicted rating
final_picks = select_genres(item_movies)#gives the app's user the choice to request films of a certain genre
print(display_films(final_picks))
e = input("I hoped those movies look good. \n Press 0 if you would like to exit or any key to start the process over:\n\n")
"288",
"1",
"300",
"121",
"174",
"127",
"56",
"7",
"98",
"237",
"117",
"172",
"222",
"204",
)
df = build_dataframe(user_dict)
def build_evaluation_file(user_dict, k):
"""Takes dict input of user ratings and returns csv of to calculate error"""
try:
new_df = pd.DataFrame(list(user_dict.items()), columns=["Movie ID", "Ratings"])
user_tr = new_df.set_index(["Movie ID"]) # Application user's movie ratings in data frame
# Create transposed versions of these dataframes in order to use the correlation method available in Pandas
# df is the queried dataframe built using the build_dataframe function in line 35
mean_df = df.fillna(df.mean()) # fill missing ratings with the mean rating across other users for each film.
df_tr = mean_df.transpose()
# Create objects to be used to calculate correlations