def prediction(movie_id):
similarities =[]
keys_list = mc.get("keys")
for key in keys_list:
sim = pearson_similarity(RATINGS, key, movie_id)
similarities.append((sim, key))
similarities.sort()
best_guess_tup = similarities[-1]
bg_sim = best_guess_tup[0]
bg_movie = best_guess_tup[1]
rating = mc.get(bg_movie)
return "Our best guess for movie %s: is %s stars" % (movie_id, rating)
def predict(movie_id):
movie_name1 = movie_dict[str(movie_id)]["title"]
pearson = []
for k, v in USER_RATING.iteritems():
movie_name2 = movie_dict[k]["title"]
pearson.append((correlation.pearson_similarity(user_ratings, movie_name1, movie_name2), v))
# get the max pearson score and assign the associated rating to prediction
sorted_list = sorted(pearson)
sorted_list.reverse()
print sorted_list
score = sorted_list[0][0]
most_similar_rating = sorted_list[0][1]
if score <= 0:
print "Cannot predict score. Please rate more movies"
return
prediction = "%.1f" % round(score * most_similar_rating)
print "Best Pearson guess for movie %s: %s is %s stars" % (movie_id, movie_name1, prediction)
def predict(target_movie_id, rating_dictionary, movie_dictionary):
movie = movie_dictionary.get(target_movie_id)
title = movie["title"]
best_key = None
best_correlation = -9
# for movie_id, rating in g_crud.items():
for movie_id in g_crud.get("my_keys"):
rating = g_crud.get(movie_id)
similarity = correlation.pearson_similarity(rating_dictionary, target_movie_id, movie_id)
if similarity > best_correlation:
best_key = movie_id
best_correlation = similarity
predicted_rating = similarity * rating
print "Best guess for movie %s: %s is %.1f stars" % (movie_id, title, predicted_rating)
def predict(movie_id):
"""Returns a prediction of movie rating based on other movies
user has rated using given movie id. If YOU_DIC is not populated
with movie ratings, an error is returned. """
movie_ratings = make_movie_ratings()
movie_title = MOVIE_DIC[movie_id]['title']
a = 0
d = []
for each in YOU_DIC:
similarity = pearson_similarity(movie_ratings, movie_title, MOVIE_DIC[each]['title'])
d.append([similarity, each]) # each is {movie_id : rating}
d.sort(reverse=True)
best = d[0]
if d[0][0] < 0.25:
return "Please rate more movies so we know what you like, then try again"
best_id = best[1]
best_rating = YOU_DIC[best_id]
return "Best guess for movie %s: %s is %s stars" %(movie_id, movie_title, best_rating)
def predict(movie_id):
ratings = RATINGS_DB[movie_id] #assigns variable 'ratings' to value of RATINGS_DB[movie_id]
movie = MOVIE_DB[movie_id] #assigns variable 'movie' to MOVIE_DB[movie_id]
movies = [ m for k, m in USER_RATINGS.iteritems() ] # THIS LINE IS DUM - christian zeb f
similarities = [ (correlation.pearson_similarity(RATINGS_DB, movie_id, k), m) for k, m in USER_RATINGS.iteritems() ]
# comparing your rating to all other ratings for these movies, and compares similarity
# tuple for movie id and similarity score
top_five = sorted(similarities) # all similar movies, sorted
top_five.reverse() # highest to lowest
top_five = top_five[:5] # actual top 5
num = 0.0
den = 0.0
# Use a weighted mean rather than a strict top similarity
for sim, m in top_five:
num += (float(sim) * m) # sim
den += sim
rating = num/den # calculated based on top5 -- the closer to 5, the more likely the movie ratings are similar to your own
print "Best guess for movie %d: %s is %.2f stars"%\
(movie_id, movie['title'], rating)