# create a list of metric classes to be evaluated
metric = []
metric.append(ac.HitRate(20))
metric.append(ac.HitRate(10))
metric.append(ac.HitRate(5))
metric.append(ac.HitRate(3))
metric.append(ac.MRR(20))
metric.append(ac.MRR(10))
metric.append(ac.MRR(5))
metric.append(ac.MRR(3))
metric.append(cov.Coverage(20))
metric.append(cov.Coverage(10))
metric.append(cov.Coverage(5))
metric.append(cov.Coverage(3))
metric.append(pop.Popularity(20))
metric.append(pop.Popularity(10))
metric.append(pop.Popularity(5))
metric.append(pop.Popularity(3))
# metric.append( div.ArtistDiversity(20) )
# metric.append( coh.ArtistCoherence(20) )
# create a dict of (textual algorithm description => class) to be evaluated
algs = {}
#baselines
mk = markov.MarkovModel()
algs['markov'] = mk
sra = sr.SequentialRules(10, weighting='div', pruning=20, last_n_days=None)
drop = random.choice(drop_outs)
momentum = random.choice(momentums)
loss = random.choice(losses)
export_csv = export_csv_base + str(test_num) + '.csv'
# create a list of metric classes to be evaluated
metric = []
metric.append(ac.HitRate(20))
metric.append(ac.HitRate(10))
metric.append(ac.HitRate(3))
metric.append(ac.MRR(20))
metric.append(ac.MRR(10))
metric.append(ac.MRR(3))
metric.append(cov.Coverage(20))
metric.append(pop.Popularity(20))
# create a dict of (textual algorithm description => class) to be evaluated
algs = {}
key = 'gru4rec2-' + loss + '-lr' + str(lr) + '-do' + str(
drop) + '-mom' + str(momentum) + 't' + str(test_num)
print('TESTING: ' + key)
drop = float(drop)
gru = gru4rec2.GRU4Rec(loss=loss,
final_act='linear',
hidden_act='tanh',
layers=[100],
batch_size=64,
dropout_p_hidden=drop,