def test_store_als(store_cls):
algo = BiasedMF(10)
algo.fit(lktu.ml_test.ratings)
with store_cls() as store:
k = store.put_model(algo)
client = store.client()
a2 = client.get_model(k)
assert a2 is not algo
assert a2.item_features_ is not algo.item_features_
assert np.all(a2.item_features_ == algo.item_features_)
assert a2.user_features_ is not algo.user_features_
assert np.all(a2.user_features_ == algo.user_features_)
del a2
def test_store_als():
algo = BiasedMF(10)
algo.fit(lktu.ml_test.ratings)
shared = lks.persist(algo)
k2 = pickle.loads(pickle.dumps(shared))
try:
a2 = k2.get()
assert a2 is not algo
assert a2.item_features_ is not algo.item_features_
assert np.all(a2.item_features_ == algo.item_features_)
assert a2.user_features_ is not algo.user_features_
assert np.all(a2.user_features_ == algo.user_features_)
del a2
k2.close()
del k2
finally:
shared.close()
def test_als_isolate(ml20m, rng):
users = rng.choice(ml20m['user'].unique(), 5000, replace=False)
algo = BiasedMF(20, iterations=10)
algo = Recommender.adapt(algo)
_log.info('training %s', algo)
ares = batch.train_isolated(algo, ml20m)
try:
_log.info('recommending with %s', algo)
recs = batch.recommend(ares, users, 10)
assert recs['user'].nunique() == 5000
_log.info('predicting with %s', algo)
pairs = ml20m.sample(1000)
preds = batch.predict(ares, pairs)
assert len(preds) == len(pairs)
finally:
ares.close()
def instantiate(opts, implicit):
feats, ureg, ireg, damp = opts
return GenderCalibratedRecommender(
BiasedMF(feats, reg=(ureg, ireg), damping=damp), getBookGender(), 0.5)
"""
Code for a gender balanced explicit ALS algorithm.
"""
from skopt.space import Integer, Real
from lenskit.algorithms.als import BiasedMF
from bookgender.rerank.GenderCalibratedRecommender import GenderCalibratedRecommender
from bookgender.rerank.fastForceGenderBalanceRecommender import FastForceGenderBalanceRecommender
from bookgender.rerank.rerankUtil import getBookGender
from bookgender.rerank.slowForceGenderBalanceRecommender import SlowForceGenderBalanceRecommender
from bookgender.rerank.slowForceGenderTargetRecommender import SlowForceGenderTargetRecommender
dimensions = [Integer(5, 300), Real(1.0e-6, 5), Real(1.0e-6, 5), Real(0, 25)]
def instantiate(opts, implicit):
feats, ureg, ireg, damp = opts
return GenderCalibratedRecommender(
BiasedMF(feats, reg=(ureg, ireg), damping=damp), getBookGender(), 0.5)
#return SlowForceGenderTargetRecommender(BiasedMF(feats, reg=(ureg, ireg), damping=damp), getBookGender())
update = None
sweep_points = [BiasedMF(nf) for nf in range(25, 250, 25)]
sweep_attrs = ['features']
import pickle
from lenskit.algorithms.als import BiasedMF
from lenskit.batch import predict
from lenskit.metrics.predict import rmse
test_bool = True
train = pd.read_pickle("../data/ml-1m-split/train.pkl")
val = pd.read_pickle("../data/ml-1m-split/val.pkl")
test = pd.read_pickle("../data/ml-1m-split/test.pkl")
num_factors = 30
num_iters = 100
model = BiasedMF(num_factors, iterations=num_iters)
print("Fitting model...")
model.fit(train)
print("Making validation predictions...")
val_preds = predict(model, val)
val_result = rmse(val_preds["prediction"], val_preds["rating"])
if test_bool:
print("Making test predictions...")
test_preds = predict(model, test)
test_result = rmse(test_preds["prediction"], test_preds["rating"])
else:
test_result = 0
print("============= RESULTS =============\nFactors: {}\nIterations: {}\nValidation RMSE: {}\nTest RMSE: {}" \
.format(num_factors, num_iters, val_result, test_result))
def default(implicit):
return BiasedMF(50)
def instantiate(opts, implicit):
feats, ureg, ireg, damp = opts
return BiasedMF(feats, reg=(ureg, ireg), damping=damp)