def __init__(self,
urm,
ucm,
binary=False,
verbose=True,
mode='offline',
datareader=None,
verbose_evaluation=True,
bm25=False,
similarity='tversky'):
assert (mode in ('offline', 'online'))
if binary: ucm.data = np.ones(ucm.data.shape[0])
self.urm = urm
self.binary = binary
self.verbose = verbose
self.verbose_ev = verbose_evaluation
self.dr = datareader
self.mode = mode
self.similarity = similarity
self.bm25 = bm25
ucm_aux = ucm.copy()
ut.inplace_set_rows_zero(X=ucm_aux,
target_rows=self.dr.get_test_pids()
) #don't learn from challange set
ucm_aux.eliminate_zeros()
if self.bm25: self.m_ui = bm25_row(ucm.copy()).tocsr()
else: self.m_ui = ucm.copy().tocsr()
if self.bm25: self.m_iu = bm25_col(ucm_aux.T.copy()).tocsr()
else: self.m_iu = ucm_aux.T.copy().tocsr()
if mode == 'offline':
self.ev = Evaluator(self.dr)
def position_matrix_boost(pos_matrix, target_rows=None, mode='linear',p1=None,p2=None,v1=None,v2=None,n_steps=None,verbose=True):
pos_matrix = sp.csr_matrix(pos_matrix, copy=True, dtype=np.float)
if target_rows is None: target_rows=np.arange(0,pos_matrix.shape[0],1)
data = []
cols = []
rows = []
for id_row in tqdm(target_rows, disable=not verbose):
row = pos_matrix[id_row]
if mode == 'linear': new_data = linear(row.data,p1,p2,v1,v2)
elif mode == 'steps': new_data = steps(row.data,n_steps,v1,v2)
else: print('MODE string error')
cols.extend(row.indices.tolist())
rows.extend(np.full(row.indices.shape[0], id_row).tolist())
data.extend(new_data)
m1 = sp.csr_matrix((data,(rows,cols)),shape=pos_matrix.shape)
ut.inplace_set_rows_zero(X=pos_matrix, target_rows=target_rows)
return pos_matrix+m1
def __init__(self,
urm,
pop=None,
binary=False,
K1=1.2,
B=0.75,
verbose=True,
mode='offline',
datareader=None,
verbose_evaluation=True,
mode_t=False,
trick=False):
assert (mode in ('offline', 'online'))
if binary: urm.data = np.ones(urm.data.shape[0])
if pop is None: self.pop = urm.sum(axis=0).A1
else: self.pop = pop
self.dr = datareader
self.urm = urm
urm_aux = urm.copy()
ut.inplace_set_rows_zero(X=urm_aux,
target_rows=self.dr.get_test_pids()
) #don't learn from challange set
urm_aux.eliminate_zeros()
if mode_t: self.m_ui = urm_aux.copy().tocsr()
else: self.m_ui = pre.bm25_row(urm_aux.copy(), K1=K1, B=B).tocsr()
if mode_t: self.m_iu = urm_aux.T.copy().tocsr()
else: self.m_iu = pre.bm25_row(urm_aux.T.copy(), K1=K1, B=B).tocsr()
self.binary = binary
self.verbose = verbose
self.verbose_ev = verbose_evaluation
self.mode = mode
self.mode_t = mode_t
if trick:
self.urm = pre.bm25_row(
urm).tocsr() #high click, high ndcg, better no use
if mode == 'offline':
self.ev = Evaluator(self.dr)
def __init__(self,
urm,
pop=None,
binary=False,
verbose=True,
mode='offline',
datareader=None,
verbose_evaluation=True,
similarity='tversky'):
assert (mode in ('offline', 'online'))
if binary: urm.data = np.ones(urm.data.shape[0])
if pop is None:
self.pop = urm.sum(axis=1).A1 #####maybe better use followers
else:
self.pop = pop
self.urm = urm
self.binary = binary
self.verbose = verbose
self.verbose_ev = verbose_evaluation
self.dr = datareader
self.mode = mode
self.similarity = similarity
urm_aux = urm.copy()
ut.inplace_set_rows_zero(X=urm_aux,
target_rows=self.dr.get_test_pids()
) #don't learn from challange set
urm_aux.eliminate_zeros()
if self.similarity == 'p3alpha':
self.m_ui = pre.bm25_col(urm.copy()).tocsr()
else:
self.m_ui = urm.copy().tocsr()
if self.similarity == 'p3alpha':
self.m_iu = pre.bm25_col(urm_aux.T.copy()).tocsr()
else:
self.m_iu = urm_aux.T.copy().tocsr()
if mode == 'offline':
self.ev = Evaluator(self.dr)
cluster_4_artist = np.array(cluster_4_artist).ravel()
cluster_1_artist_skip_lines = np.setdiff1d(all_lines, cluster_1_artist)
cluster_2_artist_skip_lines = np.setdiff1d(all_lines, cluster_2_artist)
cluster_3_artist_skip_lines = np.setdiff1d(all_lines, cluster_3_artist)
cluster_4_artist_skip_lines = np.setdiff1d(all_lines, cluster_4_artist)
## writing the clustered matrices
for path_eurm in filenames:
eurm = sps.load_npz(path_eurm)
print(path_eurm)
eurm_cluster_1_art = csr_matrix(eurm.copy())
inplace_set_rows_zero(eurm_cluster_1_art, cluster_1_artist_skip_lines)
sps.save_npz(ar1 + path_eurm.split('/')[-1],
eurm_cluster_1_art.tocsr())
eurm_cluster_2_art = csr_matrix(eurm.copy())
inplace_set_rows_zero(eurm_cluster_2_art, cluster_2_artist_skip_lines)
sps.save_npz(ar2 + path_eurm.split('/')[-1],
eurm_cluster_2_art.tocsr())
eurm_cluster_3_art = csr_matrix(eurm.copy())
inplace_set_rows_zero(eurm_cluster_3_art, cluster_3_artist_skip_lines)
sps.save_npz(ar3 + path_eurm.split('/')[-1],
eurm_cluster_3_art.tocsr())
eurm_cluster_4_art = csr_matrix(eurm.copy())
inplace_set_rows_zero(eurm_cluster_4_art, cluster_4_artist_skip_lines)
### Submission ###
#Data initialization
dr = Datareader(verbose=False, mode=mode, only_load=True)
#Recommender algorithm initialization
rec = R_p_3_beta()
#Submitter initialization
sb = Submitter(dr)
#Getting data ready for the recommender algorithm
urm = dr.get_urm()
pids = dr.get_test_pids()
urm.data = np.ones(len(urm.data))
ut.inplace_set_rows_zero(
X=urm, target_rows=pids) #don't learn from challange set
urm.eliminate_zeros()
p_ui = normalize(urm, norm="l1")
p_iu = normalize(urm.T, norm="l1")
top = urm.sum(axis=0).A1
# Fitting data
rec.fit(p_ui, p_iu, top, pids)
#Computing similarity/model
rec.compute_model(top_k=knn,
shrink=250,
alpha=0.5,
beta=0.5,
verbose=True)