def _calculateSimilarity(self, threshold=0):
""" similar_movies: (movie_id, movie_rating_rdd)"""
user_cnt = self.utility.numRows()
movie_cnt = self.utility.numCols()
rdd = self.utility.entries
sims = dict()
sims_set = dict()
for i in range(1, user_cnt + 1):
sims_set[i] = list()
users = rdd.groupBy(lambda x: x.i).collect()
for i in range(1, user_cnt + 1):
for j in range(1, user_cnt + 1):
if i == j:
continue
for user in users:
if user[0] == i:
user1 = user
if user[0] == j:
user2 = user
vt1 = SparseVector(movie_cnt,
[(user.j, 1) for user in user1[1]])
vt2 = SparseVector(movie_cnt,
[(user.j, 1) for user in user2[1]])
sim = vt1.dot(vt2) / (vt1.norm(2) * vt2.norm(2))
sims[(i, j)] = sim
sims_set[i].append(j)
return (sims, sims_set)
