xgb_hist_clf_pipeline = xgb.train(params, dtrain, num_boost_round=num_rounds)
with Timer() as t_test:
y_prob_xgb_hist = xgb_hist_clf_pipeline.predict(dtest)
# In[14]:
y_pred_xgb_hist = binarize_prediction(y_prob_xgb_hist)
# In[15]:
report_xgb_hist = classification_metrics_binary(y_test, y_pred_xgb_hist)
report2_xgb_hist = classification_metrics_binary_prob(y_test, y_prob_xgb_hist)
report_xgb_hist.update(report2_xgb_hist)
# In[16]:
results_dict['xgb_hist']={
'train_time': t_train.interval,
'test_time': t_test.interval,
'performance': report_xgb_hist
}
# In[17]:
# In general terms, leaf-wise algorithms are more efficient, they converge much faster than depth-wise. However, it may cause over-fitting when the data is small or there are too many leaves. # ### Metrics # We are going to obtain some metrics to evaluate the performance of each of the models. # ```python # report_xgb = classification_metrics_binary(y_test, y_pred_xgb) # report2_xgb = classification_metrics_binary_prob(y_test, y_prob_xgb) # report_xgb.update(report2_xgb) # results_dict['xgb']['performance'] = report_xgb # ``` # In[35]: report_xgb_hist = classification_metrics_binary(y_test, y_pred_xgb_hist) report2_xgb_hist = classification_metrics_binary_prob(y_test, y_prob_xgb_hist) report_xgb_hist.update(report2_xgb_hist) # In[36]: results_dict['xgb_hist']['performance'] = report_xgb_hist # In[37]: report_lgbm = classification_metrics_binary(y_test, y_pred_lgbm) report2_lgbm = classification_metrics_binary_prob(y_test, y_prob_lgbm)