def naive_CBF():
dm = build_data.build_urm()
target = build_data.loadTarget()
rec = r.CBF_coldstart(k=50)
rec.fit()
recommended = rec.recommendAll(target, 10)
playlists = recommended[:, 0]
recommended = numpy.delete(recommended, 0, 1)
i = 0
res_fin = []
for j in recommended:
res = ''
for k in range(0, len(j)):
res = res + '{0} '.format(j[k])
res_fin.append(res)
i = i + 1
d = {'playlist_id': playlists, 'track_ids': res_fin}
df = pandas.DataFrame(data=d, index=None)
df.to_csv("../results/recommended9.csv", index=None)
i = 1 + 1
def __init__(self, features):
self._lambda = 0.1
self._alpha = 40
self._URM = bd.build_urm().tocsr()
self._users = self._URM.shape[0]
self._items = self._URM.shape[1]
self._X = sps.csr_matrix(np.random.normal(size=(self._users,
features)))
self._Y = sps.csr_matrix(np.random.normal(size=(self._items,
features)))
self._X_I = sps.eye(self._users)
self._Y_I = sps.eye(self._items)
self._I = sps.eye(features)
self._II = self._I * self._lambda
self._xTx = self._X.T.dot(self._X)
self._Cui = self._URM * self._alpha
self._ALS_ITERS = 100
def __init__(self, k, coldstart=10):
self._coldstart = coldstart
self._cbf = CBF_Item_Naive(k)
self._URM = bd.build_urm().tocsr()
self._toppop = TopPop()
def __init__(self, k):
self._cosine = sim.Cosine_Similarity(bd.build_icm(), k)
self.URM = bd.build_urm()
self._k = k