predictions = rs.recommend(target_data['playlist_id'])
map_ = (evaluator.evaluate(predictions, test_data))
map_list.append(map_)
df = df.append(
pd.DataFrame([[k, map_, shrinkage]], columns=['knn', 'map',
'shr']))
top_50 = df.sort_values(by=['map']).tail(50)
knn_list.append(k)
k += 10
print(top_50)
if plot_graph:
plt.plot(knn_list, map_list, 'bs')
plt.title(shrinkage)
plt.show()
save_dataframe('../output/content_w_tuning_df.csv', ',', df)
shrinkage += 10
print(top_50)
print('End of parameter tuning')
'''
from cbfRS.cbfRS import CbfRS
from loader.loader import save_dataframe, train_data, target_data, full_data, test_data, tracks_data
from utils.auxUtils import Evaluator
evaluator = Evaluator()
i = 0
list_map = []
while i < 6:
while shrinkage < 600:
map_list = []
knn_list = []
k = 100
while k <= 800:
rs = ColBfIIRS(10, k, shrinkage, tf_idf=True)
rs.fit(train_data)
print('knn: ', k, ' shrinkage: ', shrinkage)
predictions = rs.recommend(target_data['playlist_id'])
map_ = (evaluator.evaluate(predictions, test_data))
map_list.append(map_)
df = df.append(
pd.DataFrame([[k, map_, shrinkage]], columns=['knn', 'map',
'shr']))
top_50 = df.sort_values(by=['map']).tail(50)
knn_list.append(k)
k += 50
print(top_50)
shrinkage += 50
if plot_graph:
plt.plot(knn_list, map_list, 'bs')
plt.title(shrinkage)
plt.show()
save_dataframe('../output/coll_item_item_tuning_df.csv', ',', df)
print(top_50)
print("End of parameter tuning")
# Get values of both in common structure of CF
W_sparse_delta = W_sparse_CBF.copy().multiply(W_sparse_common)
W_sparse_delta -= W_sparse_CF.copy().multiply(W_sparse_common)
W_sparse_delta_sorted = np.sort(W_sparse_delta.data.copy())
'''
pyplot.plot(W_sparse_CF_sorted, label='CF')
pyplot.plot(W_sparse_CBF_sorted, label='CBF')
'''
pyplot.plot(W_sparse_delta_sorted, label='delta')
pyplot.ylabel('Similarity cell ')
pyplot.xlabel('Similarity value')
pyplot.legend()
pyplot.show()
print(
"W collaborative item item has {:.2E} values and {:.2f} % in common with CBF"
.format(W_sparse_CF.nnz, W_sparse_common.nnz / W_sparse_CF.nnz * 100))
print("W content has {:.2E} values and {:.2f} % in common with CF".format(
W_sparse_CBF.nnz, W_sparse_common.nnz / W_sparse_CBF.nnz * 100))
print("W_sparse_delta has {:.2E} values".format(W_sparse_delta.nnz))
W_sparse_delta = W_sparse_delta.tocoo()
CFW_weithing = CFW_D_Similarity_Linalg(URM_train, ICM_all, W_sparse_CF,
train_data)
CFW_weithing.fit()
prediction = CFW_weithing.recommend(target_data['playlist_id'])
evaluator.evaluate(prediction, test_data)
save_dataframe('output/hybrid_submission.csv', ',', prediction)
while shrinkage < 300:
map_list = []
knn_list = []
k = 50
while k < 500:
rs = ColBfUURS(10, k, shrinkage, tf_idf=True)
rs.fit(train_data)
print('knn: ', k, ' shrinkage: ', shrinkage)
predictions = rs.recommend(target_data['playlist_id'])
map_ = (evaluator.evaluate(predictions, test_data))
map_list.append(map_)
df = df.append(
pd.DataFrame([[k, map_, shrinkage]], columns=['knn', 'map',
'shr']))
top_50 = df.sort_values(by=['map']).tail(50)
k += 50
print(top_50)
if plot_graph:
plt.plot(knn_list, map_list, 'bs')
plt.title(shrinkage)
plt.show()
shrinkage += 25
save_dataframe('../output/coll_user_user_tuning.csv', ',', top_50)
print(top_50)
print("End of parameter tuning")
pd.DataFrame([[alpha, beta, gamma, temp_map]],
columns=['alpha', 'beta', 'gamma', 'map']))
top_10 = df.sort_values(by=['map']).tail(10)
gamma += 1
print(top_10)
beta += 1
alpha += 1
# print(df)
top_50 = df.sort_values(by=['map']).tail(50)
print("FINAL RESULTS")
print(top_50)
save_dataframe('output/hybrid_col_uu_col_ii_cbf.csv', ',', df)
'''
Old tuning:
# Hybrid (cbf - colf)
rs = HybridRS(tracks_data, 10, tf_idf=True)
evaluator = Evaluator()
rs.fit(train_data)
delta = 0.5
while delta <= 1:
slim_p = rs.recommend_probability(target_data['playlist_id'], 3, 6, 2, delta)
evaluator.evaluate(slim_p, test_data)
slim = rs.recommend(target_data['playlist_id'], 3, 6, 2, delta)
evaluator.evaluate(slim, test_data)
delta += 0.1
# print(df)
r.fit(buildURMMatrix(train_data))
pred = r.recommend(target_data['playlist_id'])
temp_map = e.evaluate(pred, test_data)
print("Alpha: ", a)
print("Reg: ", reg)
print("Num factors: ", nf)
print("MAP: ", temp_map)
'''
'''
for r in regs:
rs = IALS_numpy(num_factors=250, reg=r)
e = Evaluator()
rs.fit(buildURMMatrix(train_data))
pred = rs.recommend(target_data['playlist_id'])
temp_map = e.evaluate(pred, test_data)
print("Regularization: ", r)
print("MAP: ", temp_map)
'''
rs = IALS_numpy(num_factors=2, reg=100)
e = Evaluator()
rs.fit(buildURMMatrix(train_data))
print("GOING")
print(rs.get_estimated_ratings())
pred = rs.recommend(target_data['playlist_id'])
e.evaluate(pred, test_data)
save_dataframe('output/als_250_100_reg_factors.csv', sep=',', dataframe=pred)
from graphBased.p3alpha import P3alphaRecommender
from loader.loader import save_dataframe, train_data, target_data, full_data, test_data, tracks_data
from utils.auxUtils import Evaluator
evaluator = Evaluator()
rs = P3alphaRecommender(train_data)
# best values 1.1, topK= 75
rs.fit()
predictions = rs.recommend(target_data['playlist_id'])
map_ = (evaluator.evaluate(predictions, test_data))
save_dataframe('output/p3_alpha_tuning.csv', ',', predictions)
# from lightfm import LightFM
from cbfRS.cbfRS import CbfRS
from collaborative_filtering_RS.col_user_userRS import ColBfUURS
from collaborative_filtering_RS.col_item_itemRS import ColBfIIRS
from hybrid_col_cbf_RS.hybridRS import HybridRS
from slimRS.Cython.SLIM_BPR_Cython import SLIM_BPR_Cython
# from hybrid_col_cbf_RS.hybrid_pureSVD import HybridRS
from svdRS.pureSVD import PureSVDRecommender
import matplotlib.pyplot as plt
from slimRS.slimElasticNet import SLIMElasticNetRecommender
from hybrid_col_cbf_RS.hybrid_als_knn_slimBPR_elasticNet import HybridRS
rs = HybridRS(tracks_data)
rs.fit(full_data)
pred = rs.recommend(target_data['playlist_id'])
save_dataframe('output/hybrid_als.csv', sep=',', dataframe=pred)
''' PHI TUNNING
map_list = []
phi_list = []
phis = [8, 10, 12, 14, 16, 18, 20, 22, 24]
evaluator = Evaluator()
rs = HybridRS(tracks_data)
rs.fit(train_data)
for phi in phis:
predictions = rs.recommend(target_data['playlist_id'], phi=phi)
map_ = evaluator.evaluate(predictions, test_data)
print("PHI: ", phi)
print("MAP: ", map_)
map_list.append(map_)
phi_list.append(phi)