train_['rel'],
qids_val.values,
metric=metric)
model = LambdaMART(metric=metric,
max_depth=10,
n_estimators=1000,
learning_rate=.1,
verbose=1,
max_features=0.5,
query_subsample=0.5,
max_leaf_nodes=10,
min_samples_leaf=64)
model.fit(train_[m2_rnn], train_['rel'], qids_val, monitor=monitor)
#creating the data to test the models with
LambdaMART_testX = test_[m2_rnn].set_index('srch_id').sort_index()
LambdaMART_testY = test_[['srch_id', 'booking_bool', 'click_bool']]
LambdaMART_test_qids = LambdaMART_testX.index
#LambdaMART predictions
Lambda_pred = model.predict(LambdaMART_testX)
print('Random ranking: ',
metric.calc_mean_random(LambdaMART_test_qids, LambdaMART_testY))
print('Our model: ',
metric.calc_mean(LambdaMART_test_qids, LambdaMART_testY, Lambda_pred))
#old way, probably not correct for LambdaMART
output_LM_out = model.predict(test[m2_rnn])
result = test[['srch_id', 'booking_bool', 'click_bool']]
result = result.assign(pred=output_LM_out)
# In[5]:
model = LambdaMART(metric=metric, max_depth = 4, n_estimators=450, learning_rate=.04, verbose=1, max_features = 25,
min_samples_split = 1000, min_samples_leaf = 200, max_leaf_nodes = 20)
model.fit(x_train, y_train, qids_train, monitor=monitor)
# In[ ]:
prediction = model.predict(x_test)
print ('Random ranking:', metric.calc_mean_random(qids_test, y_test))
print ('Our model:', metric.calc_mean(np.asarray(qids_test), np.asarray(y_test), prediction))
prediction_train = model.predict(x_train)
print ('Train model:', metric.calc_mean(np.asarray(qids_train), np.asarray(y_train), prediction_train))
# In[ ]:
#a = zip(model.feature_importances_, x_train.columns)
#b = list()
#for imp, name in a:
# b.append((imp, name))
#sorted(b)