def cascade_LR(): #意义不大
if config.cascade_LR:
LinearRgressor = quantum_forest.Linear_Regressor({'cascade':"ridge"})
y_New = LinearRgressor.BeforeFit((data.X_train, data.y_train),[(data.X_valid, data.y_valid),(data.X_test, data.y_test)])
YY_train = y_New[0]
YY_valid,YY_test = y_New[1],y_New[2]
else:
YY_train,YY_valid,YY_test = data.y_train, data.y_valid, data.y_test
return YY_train,YY_valid,YY_test
def Fold_learning(fold_n,data,config,visual):
t0 = time.time()
if config.model=="QForest":
if config.feat_info == "importance":
feat_info = get_feature_info(data,fold_n)
else:
feat_info = None
accu,_ = NODE_test(data,fold_n,config,visual,feat_info)
elif config.model=="GBDT":
accu,_ = GBDT_test(data,fold_n)
else: #"LinearRegressor"
model = quantum_forest.Linear_Regressor({'cascade':"ridge"})
accu,_ = model.fit((data.X_train, data.y_train),[(data.X_test, data.y_test)])
print(f"\n======\n====== Fold_{fold_n}\tACCURACY={accu:.5f},time={time.time() - t0:.2f} ====== \n======\n")
return
def Fold_learning(fold_n, data, config, visual):
t0 = time.time()
if config.model == "QForest":
if config.feat_info == "importance":
feat_info = get_feature_info(data, fold_n)
else:
feat_info = None
accu, _ = QF_test(data, fold_n, config, visual, feat_info)
elif config.model == "GBDT" or config.model == "Catboost" or config.model == "XGBoost" or config.model == "LightGBM":
accu, _ = GBDT_test(config,
data,
fold_n,
num_rounds=config.nMostEpochs)
else: #"LinearRegressor"
model = quantum_forest.Linear_Regressor({'cascade': "ridge"})
accu, _ = model.fit((data.X_train, data.y_train),
[(data.X_test, data.y_test)])
print(f"\n======\n====== Fold_{fold_n}@{data.name}\t{data.problem()}"\
f"\tACCURACY={accu:.5f},time={time.time() - t0:.2f} ====== \n======\n")
return