def getAverageMoviesRatings(udata_array):
#summ of ratings, rating count
#iterator = movieId
array = np.zeros(shape=(MOVIE_COUNT, 2), dtype=int)
for i in range(len(udata_array)):
movieId = udata_array[i, 1]
rating = udata_array[i, 2]
normalized_rating = common.getQuantizedRating(rating)
array[movieId - 1, 1] += 1
array[movieId - 1, 0] += normalized_rating
arrayWithAverageRatings = np.zeros(shape=(MOVIE_COUNT, 1), dtype=int)
for i in range(MOVIE_COUNT):
ratingCount = array[i, 1]
ratingSumm = array[i, 0]
avg = 0.0
if (ratingCount > 0):
avg = float(ratingSumm) / ratingCount
if (avg > 0.5):
arrayWithAverageRatings[i] = 1
return arrayWithAverageRatings
def getQuantizedAverageRecomendationPoint(arrayWithAverageRatings, user_rating,
movieId):
normalized_user_rating = common.getQuantizedRating(user_rating)
if (normalized_user_rating == int(arrayWithAverageRatings[movieId - 1])):
return 1
else:
return 0
def getNormalizedAvgRatingPotinFromSimilarTaste(ratingsCorelation3, ratingsCorelation2, ratingsCorelation1, rating):
normalizedOtherRating = getNormalizedAveragedRatingFromSimilarTaste(ratingsCorelation3, ratingsCorelation2, ratingsCorelation1, rating)
normalizedRating = common.getQuantizedRating(rating)
#print(str(normalizedRating) + " " + str(normalizedOtherRating))
if(normalizedOtherRating == normalizedRating):
return 1
else:
return 0
def getNormalizedAveragedRatingFromSimilarTaste(ratingsCorelation3, ratingsCorelation2, ratingsCorelation1, rating):
bestUsersRatings = getBestFitSimilarUsersDictionary(ratingsCorelation3, ratingsCorelation2, ratingsCorelation1)
summ = 0
count = len(bestUsersRatings)
for i in range(count):
summ += bestUsersRatings[i]
avgRating = float(summ)/float(count)
return common.getQuantizedRating(avgRating)
def getRandomRSPrecisionRecall(udata_array):
count = 0
summ = 0
finalCount = len(udata_array)
precisionList = []
recallList = []
for row in udata_array:
user_rating = common.getQuantizedRating(row[2])
system_rating = common.getQuantizedRating(random.randint(1, 5))
count += 1
if(user_rating == system_rating):
summ += 1
precisionList.append(summ/count)
recallList.append(summ/finalCount)
return common.PrecisionRecallModel(precisionList, recallList)
def getNormalizedAvgRatingPotinFromSimilarTaste(ratingsCorelation3,
ratingsCorelation2,
ratingsCorelation1, rating):
normalizedOtherRating = getNormalizedAveragedRatingFromSimilarTaste(
ratingsCorelation3, ratingsCorelation2, ratingsCorelation1, rating)
normalizedRating = common.getQuantizedRating(rating)
#print(str(normalizedRating) + " " + str(normalizedOtherRating))
if (normalizedOtherRating == normalizedRating):
return 1
else:
return 0
def getRandomRSPrecisionRecall(udata_array):
count = 0
summ = 0
finalCount = len(udata_array)
precisionList = []
recallList = []
for row in udata_array:
user_rating = common.getQuantizedRating(row[2])
system_rating = common.getQuantizedRating(random.randint(1, 5))
count += 1
if (user_rating == system_rating):
summ += 1
precisionList.append(summ / count)
recallList.append(summ / finalCount)
return common.PrecisionRecallModel(precisionList, recallList)
def getNormalizedAveragedRatingFromSimilarTaste(ratingsCorelation3,
ratingsCorelation2,
ratingsCorelation1, rating):
bestUsersRatings = getBestFitSimilarUsersDictionary(
ratingsCorelation3, ratingsCorelation2, ratingsCorelation1)
summ = 0
count = len(bestUsersRatings)
for i in range(count):
summ += bestUsersRatings[i]
avgRating = float(summ) / float(count)
return common.getQuantizedRating(avgRating)
def getQuantizedList(mList):
qList = []
for i in range(len(mList)):
qList.append(common.getQuantizedRating(mList[i]))
return qList
def getQuantizedList(mList):
qList = []
for i in range(len(mList)):
qList.append(common.getQuantizedRating(mList[i]))
return qList