def test_user_knn():
ubknn = pyreclab.UserKnn(dataset='dataset/u1.base',
dlmchar=b'\t',
header=False,
usercol=0,
itemcol=1,
ratingcol=2)
ubknn.train(k=100, similarity='pearson')
pred = ubknn.predict('457', '443')
assert abs(pred - expected_prediction) < prediction_epsilon
ranking = ubknn.recommend('457', 5, includeRated=False)
assert ranking == expected_ranking
predlist, mae, rmse = ubknn.test(input_file='dataset/u1.test',
dlmchar=b'\t',
header=False,
usercol=0,
itemcol=1,
ratingcol=2,
output_file='predictions.csv')
assert abs(mae - expected_mae) < mae_epsilon
assert abs(rmse - expected_rmse) < rmse_epsilon
def main(k=100):
ubknn = pyreclab.UserKnn( dataset = 'dataset/u1.base',
dlmchar = b'\t',
header = False,
usercol = 0,
itemcol = 1,
ratingcol = 2 )
print( '-> training model' )
start = time.clock()
ubknn.train(k=k, similarity = 'pearson' )
end = time.clock()
print( 'training time: ' + str( end - start ) )
print( '-> individual test' )
pred = ubknn.predict( '457', '443' )
print( 'user 457, item 443, prediction ' + str( pred ) )
ranking = ubknn.recommend( '457', 5, includeRated = False )
print( 'recommendation for user 457: ' + str( ranking ) )
print( '-> prediction test' )
start = time.clock()
predlist, mae, rmse = ubknn.test( input_file = 'dataset/u1.test',
dlmchar = b'\t',
header = False,
usercol = 0,
itemcol = 1,
ratingcol = 2,
output_file = 'predictions.csv' )
end = time.clock()
print( 'prediction time: ' + str( end - start ) )
print( 'MAE: ' + str( mae ) )
print( 'RMSE: ' + str( rmse ) )
print( '-> recommendation test' )
start = time.clock()
recommendList = ubknn.testrec( input_file = 'dataset/u1.test',
dlmchar = b'\t',
header = False,
usercol = 0,
itemcol = 1,
ratingcol = 2,
topn = 10,
output_file = 'ranking.json',
includeRated = False )
end = time.clock()
print( 'recommendation time: ' + str( end - start ) )
return mae, rmse
prediction_filename = data_url+'predictionsUserKnn' #5.csv'
ordenada_filename = data_url+'ordenadasUserKnn' #5.csv'
print 'UserKnn'
for i in range(1, data_chunks + 1):
f_t = training_filename + str(i) + ".txt"
f_p = probe_filename + str(i) + ".txt"
f_pred = prediction_filename + str(i) + ".csv"
f_ord = ordenada_filename + str(i) + ".csv"
print "Corriendo experimento ", i, "..."
print 'Entrenando...'
obj = pyreclab.UserKnn( dataset = f_t,
dlmchar = b'\t',
header = True,
usercol = 0,
itemcol = 1,
ratingcol = 2 )
obj.train(10, 'pearson')
print 'Prediciendo...'
#prediction = obj.predict( "630685", "1")
#ranking = obj.recommend( "630685", 10, True)
#print prediction
#print ranking
predictionList, mae, rmse = obj.test( input_file = f_p,
dlmchar = b'\t',
header = False,
usercol = 0,
itemcol = 1,
ratingcol = 2,
import time
import pyreclab
if __name__ == '__main__':
ubknn = pyreclab.UserKnn(dataset='dataset/u1.base',
dlmchar=b'\t',
header=False,
usercol=0,
itemcol=1,
ratingcol=2)
print('-> training model')
start = time.clock()
ubknn.train(k=100, similarity='pearson')
end = time.clock()
print('training time: ' + str(end - start))
print('-> individual test')
pred = ubknn.predict('457', '443')
print('user 457, item 443, prediction ' + str(pred))
ranking = ubknn.recommend('457', 5, includeRated=False)
print('recommendation for user 457: ' + str(ranking))
print('-> prediction test')
start = time.clock()
predlist, mae, rmse = ubknn.test(input_file='dataset/u1.test',
dlmchar=b'\t',
header=False,
usercol=0,
help='input qual data')
parser.add_argument('--output',
'-o',
required=True,
type=str,
help='output file path')
parser.add_argument('--code_test',
default=0,
type=int,
help='test code using UserAvg')
args = parser.parse_args()
if (args.code_test == 0):
obj = pyreclab.UserKnn(dataset=args.train,
dlmchar=b' ',
header=False,
usercol=0,
itemcol=1,
ratingcol=3)
else:
obj = pyreclab.UserAvg(dataset=args.train,
dlmchar=b' ',
header=False,
usercol=0,
itemcol=1,
ratingcol=3)
# Train
print("Training starting")
knn = 100
similarity = 'pearson'
