def get_recommendation(self, u, b):
s = utils.scores_sorter(self.model[u].T[0])
rac = []
j = 0
while (len(rac) < 500):
if (s[j] not in self.test_users[u][:b]):
rac.append(self.items_list[s[j]])
j += 1
return rac
def build_selection(k, p2s_test, global_popularity, t2t_id, S, p2t, lengths,
test_seed, P, power):
selection = {}
for user in p2s_test[test_seed]:
if (p2t[str(user)] == "" or p2t[str(user)] not in t2t_id):
selection[user] = global_popularity[:k]
else:
row = S[t2t_id[p2t[str(user)]], :]**power / lengths
selection[user] = utils.scores_sorter(P.dot(row), k)
print("selection built")
utils.fl()
return selection
def get_recommendation(self, u, b):
s = utils.scores_sorter(self.model[u])
rac = []
j = 0
while (len(rac) < 500):
if (s[j] not in self.test_users[u][:b]):
rac.append(self.items_list[s[j]])
j += 1
return rac
def get_selection(self, u, b, k):
s = utils.scores_sorter(self.model[u], k + b)
rac = [i for i in s if i not in self.test_users[u][:b]][:k]
return rac
def get_selection(self, u, b, k):
s = utils.scores_sorter(self.model[u], k + b)
rac = [i for i in s if i not in self.test_users[u][:b]][:k]
return rac
row.append(utr2utr_id[user])
col.append(itr2itr_id[song])
data += data_row
R = sps.csr_matrix((data, (row, col)))
print("R built with shape: " + str(R.shape))
utils.fl()
itr_id2itr = {itr2itr_id[s]: s for s in itr2itr_id}
Rhat = (Ku.power(q).dot(R))
print(Rhat.shape)
recommendation = {}
for user in u2u_id:
print(user)
scores = Rhat[u2u_id[user]].toarray()[0]
idx = utils.scores_sorter(scores, 500)
recommendation[str(user)] = [itr_id2itr[i] for i in idx if scores[i] > 0]
recommendation[str(user)] = [
s_id2spt[i] for i in recommendation[str(user)]
if i not in p2s_test[test_seed][user][:int(test_seed)]
][:500]
k = 0
while len(recommendation[str(user)]) < 500:
track = s_id2spt[global_popularity[k]]
if track not in recommendation[str(user)] and global_popularity[
k] not in p2s_test[test_seed][user][:int(test_seed)]:
recommendation[str(user)].append(track)
k += 1
utils.save_recommendation("./1.csv", recommendation)
p2s_test = utils_matrix.load_test_set("./")
s_id2spt = utils.jload("./s_id2spt.json")
print "loading P...",
utils.fl()
P = sps.load_npz("./P.npz")
print "done"
utils.fl()
print "loading P csc...",
utils.fl()
P_csc = sps.load_npz("./P_csc.npz")
print "done"
utils.fl()
print("prediction with alpha=%f, q=%f started" % (alpha, q))
recommendation = {}
for u in sorted(p2s_test[test_seed].keys()):
Iu = p2s_test[test_seed][u][:int(test_seed)]
Wu = (P_csc[:, Iu].power(alpha).multiply(
P[Iu, :].power(1 - alpha).transpose())).power(q)
scores = np.array(np.sum(Wu, axis=1)).flatten()
recommendation[u] = [
s_id2spt[i] for i in utils.scores_sorter(scores) if i not in Iu
][:500]
print(u)
utils.fl()
#here you should specify the directory where to save the recommendation
utils.save_recommendation("./%s.csv" % (test_seed), recommendation)
data = []
col = []
row = []
for t in w2pop:
for s in w2pop[t]:
data.append(w2pop[t][s])
col.append(int(s))
row.append(t2t_id[t])
P = sps.csr_matrix((data, (col, row))) #the matrix needs to be transposed
global_popularity = np.zeros(len(s2p))
for s in s2p:
global_popularity[int(s)] = len(s2p[s])
global_popularity = np.argsort(-global_popularity)
recommendation = {}
for user in p2s_test[bucket]:
if p2t[str(user)] in t2t_id:
row = S[t2t_id[p2t[str(user)]], :]**q
recommendation[user] = [
s_id2spt[s] for s in utils.scores_sorter(P.dot(row))
][:500]
else:
recommendation[user] = [s_id2spt[s] for s in global_popularity[:500]]
#here you should specify the directory where to save the recommendation
utils.save_recommendation("./" + str(bucket) + ".csv",
recommendation,
intestation=True)