def invoke_xgb(data_train, data_test, params):
dtrain = survival_dmat(data_train, t_col=col_t, e_col=col_e, label_col="Y")
dtest = survival_dmat(data_test, t_col=col_t, e_col=col_e, label_col="Y")
# params
params_model = {
'eta': params['eta'],
'max_depth': params['max_depth'],
'min_child_weight': params['min_child_weight'],
'subsample': params['subsample'],
'colsample_bytree': params['colsample_bytree'],
'lambda': params['reg_lambda'],
'gamma': params['reg_gamma'],
'silent': 1
}
# Build and train model
model = BecCox(
params_model,
loss_alpha=params["loss_alpha"]
)
eval_result = model.train(
dtrain,
num_rounds=params['nrounds'],
silent=True,
plot=False
)
# Evaluation
return model.evals(dtest)
def invoke_xgb(data_train, data_test, params):
#print params
dtrain = survival_dmat(data_train, t_col=T_col, e_col=E_col, label_col="Y")
dtest = survival_dmat(data_test, t_col=T_col, e_col=E_col, label_col="Y")
params_xgb = {
'eta': params['eta'],
'max_depth': params['max_depth'],
'min_child_weight': params['min_child_weight'],
'subsample': params['subsample'],
'colsample_bytree': params['colsample_bytree'],
'lambda': params['reg_lambda'],
'gamma': params['reg_gamma'],
'objective': 'multi:softprob',
'num_class': K + 1,
'silent': 1,
'seed': 42
}
# Build and train model
model = HitBoost(params_xgb,
loss_alpha=params['loss_alpha'],
loss_gamma=params['loss_gamma'])
eval_result = model.train(dtrain,
num_rounds=params['nrounds'],
silent=True,
plot=False)
# Evaluation
return model.evals(dtest)