class XGB(BaseModel):
def __init__(self):
self.clf = XGBClassifier(
n_estimators=200,
max_depth=20,
learning_rate=0.1,
random_state=0,
booster="gbtree",
use_label_encoder=False,
)
def train(self, X_train, Y_train):
X_train, Y_train = do_rebalance(X_train, Y_train)
self.clf.fit(X_train, Y_train)
def test(self, X_test, Y_test):
Y_prob = self.clf.predict_proba(X_test)
auc = metrics.roc_auc_score(Y_test, Y_prob[:, 1])
def predict(self, X):
Y_prob = self.clf.predict_proba(X)
return Y_prob
def load_model(self, model_path):
self.clf.load_model(model_path)
# with open(model_path, "rb+") as file:
# self.clf = pickle.load(file)
def save_model(self, model_path):
self.clf.save_model(model_path)
class WrappedXGBClassifier(WrappedModel):
def base_init_finished(self):
self.reset()
def fit(self, X, y):
self._value.fit(X, y, **self._fit_kwargs)
return self
def reset(self):
from xgboost import XGBClassifier
self._value = XGBClassifier(**self._init_kwargs)
def predict(self, X):
return self._value.predict(X)
def predict_proba(self, X):
if self._pos_index is None:
raise Exception('predict_proba need pos_index')
return self._value.predict_proba(X)[:, self._pos_index]
def dump(self, dirpath, name):
self.value.save_model(pathjoin(dirpath, name + '.bin'))
return self
def load(self, dirpath, name):
self._value.load_model(pathjoin(dirpath, name + '.bin'))
return self
class XGBoost_Ranker():
def __init__(self, timestamp, load=True):
self.model = XGBClassifier()
self.model.load_model(timestamp + '.file')
self.factor = 1.0
def set_factor(self, factor):
self.factor = factor
def rank_features(self, features):
_features = np.copy(features)
for f in _features:
f[1] *= self.factor
f[4] *= self.factor
f[5] *= self.factor
# return np.array([0, 1, 2, 3, 4])
test_x = []
for i in range(len(_features)):
for j in range(len(_features)):
if i == j:
continue
test_x.append(
np.concatenate((_features[i], _features[j]), axis=0))
test_x = np.array(test_x)
print(test_x.shape)
y = self.model.predict(test_x).reshape(len(_features),
len(_features) - 1)
y = np.sum(y, axis=1)
# print(y)
return np.argsort(y)[::-1]
class StabilityClassifier():
def __init__(self, modelfile='spock.json'):
pwd = os.path.dirname(__file__)
self.model = XGBClassifier()
self.model.load_model(pwd + '/models/'+modelfile)
def check_errors(self, sim):
if sim.N_real < 4:
raise AttributeError("SPOCK Error: SPOCK only applicable to systems with 3 or more planets")
def predict_stable(self, sim):
triofeatures, stable = self.generate_features(sim)
if stable == False:
return 0
trioprobs = self.predict_from_features(triofeatures)
return trioprobs.min() # minimum prob among all trios tested
def generate_features(self, sim):
sim = sim.copy()
init_sim_parameters(sim)
self.check_errors(sim)
trios = [[i,i+1,i+2] for i in range(1,sim.N_real-2)] # list of adjacent trios
featureargs = [10000, 80, trios]
triofeatures, stable = features(sim, featureargs)
return triofeatures, stable
def predict_from_features(self, triofeatures):
# xgboost model expects a 2D array of shape (Npred, Nfeatures) where Npred is number of samples to predict, Nfeatures is # of features per sample
featurevals = np.array([[val for val in features.values()] for features in triofeatures])
return self.model.predict_proba(featurevals)[:,1] # take 2nd column for probability it belongs to stable class
def load_modele(path):
'''Renvoie le modele en tant qu\'objet à partir du chemin'''
if 'GradientBoosting' in str(path):
#print('Chargement XGBOOST')
model = XGBClassifier()
model.load_model(path)
return model
else:
#print('Chargement Pickle')
return pickle.load(open(path, 'rb'))
def export_model(amnt_data, client):
model = XGBClassifier()
model.load_model('boa.model')
output = open('model.pb', 'wb')
pickle.dump([model, amnt_data], output)
output.close()
bucket_test = client.get_bucket('traina-data')
blob_test = bucket_test.blob('model.pb')
blob_test.upload_from_filename(filename='model.pb')
os.remove('model.pb')
print(Fore.GREEN+'Exported Model Sucessfully')
return
def get_model():
param_path = os.path.join(STORAGE, "params.json")
with open(param_path, "r") as f:
json_data = f.read()
params = dict(json.loads(json_data))
model = XGBClassifier(**params)
model_path = os.path.join(STORAGE, "model.xgb")
model.load_model(model_path)
return model
class ProcessPlugin(WorkerPlugin):
def __init__(self,
cfg_path=os.environ.get("MODEL_CONFIG"),
weights_path=os.environ.get("PP_WEIGHTS_PTH"),
classes_path=os.environ.get("CLASSES_PTH")):
self.cfg_path = cfg_path
postprocess_weights_pth = weights_path
self.postprocess_model = XGBClassifier()
self.postprocess_model.load_model(postprocess_weights_pth)
self.classes_pth = classes_path
with open(self.classes_pth) as stream:
self.classes = yaml.load(stream)["CLASSES"]
def train_lazy():
# Load the dataset
X, y = load_data()
# Split the data
X_train, X_val, y_train, y_val = split_dataset(X, y)
# # Normalize
X_train = normalize(X_train)
X_val = normalize(X_val)
# uncomment to check the performance of the 25 models
# clf = LazyClassifier(verbose=0,ignore_warnings=True, custom_metric=None)
# # fit
# scores,_ = clf.fit(X_train, X_val, y_train, y_val)
# # print
# print(scores)
# Final model
# check if model exist
if os.path.isfile(config.MODEL_PATH):
model = XGBClassifier()
model.load_model(config.MODEL_PATH)
else:
model = XGBClassifier()
model.fit(X_train,
y_train,
eval_metric="error",
eval_set=[(X_train, y_train), (X_val, y_val)],
verbose=True)
# save model
model.save_model(config.MODEL_PATH)
# performance on train set
y_pred = model.predict(X_train)
# evaluate predictions
print_performance(y_train, y_pred, 'train')
# performance on val set
y_pred = model.predict(X_val)
# evaluate predictions
print_performance(y_val, y_pred, 'val')
# Load the test dataset
X_test, y_test = load_test_data()
# # Normalize
X_test = normalize(X_test)
# get prediction
y_pred = model.predict(X_test)
# evaluate predictions
print_performance(y_test, y_pred, 'test')
# print
plot_performance(model)
def predict_probability_of_winning(gold_diff_at_10, exp_diff_at_10, team):
dirname = os.path.dirname(__file__)
os.path.join(dirname, 'djangoapp/chart/utils/models/red_model.json')
model = XGBClassifier()
model.load_model(os.path.join(dirname, f'./models/{team}_model.json'))
df = pd.DataFrame({
team + 'GoldDiff': [gold_diff_at_10],
team + 'ExperienceDiff': [exp_diff_at_10]
})
# values are casted into lists because pandas constructor doesnt allow scalars
predicts = model.predict_proba(df)
for i, col in enumerate(['redWin', 'blueWin']):
df[col] = predicts[:, i]
return df
def generate_shap_html(feature, user_bin, user_id):
xgb_clf = XGBClassifier()
xgb_clf.load_model(os.path.join(MODEL_DIRECTORY, "xgb.model"))
explainer = shap.TreeExplainer(xgb_clf)
values = explainer.shap_values(feature)
shap.initjs()
fp = shap.force_plot(explainer.expected_value[user_bin - 1],
values[user_bin - 1][0],
feature,
show=False)
shap.save_html(os.path.join(MODEL_DIRECTORY, f"User_{user_id}.html"), fp)
with open(os.path.join(MODEL_DIRECTORY, f"User_{user_id}.html"),
"r",
encoding='utf-8') as f:
html = f.read()
os.remove(os.path.join(MODEL_DIRECTORY, f"User_{user_id}.html"))
return str(html), values
def load_model_and_generate_evaluation_images(*, model_filename, input_path,
output_path, feature_names):
model = XGBClassifier()
model.load_model(model_filename)
frame_folders = sorted(get_frame_folders(input_path))
for frame_folder in frame_folders:
frame_path = os.path.join(input_path, frame_folder)
segment_names = [
name for name in os.listdir(frame_path) if name[1].isdigit()
]
if len(segment_names) != 0:
continue
for camera_name in ["60", "180", "300"]:
image_name = "camera" + camera_name + ".png"
print(frame_path + "/" + image_name)
image_bgr = cv.imread(os.path.join(frame_path, image_name))
features, shape = create_features(image_bgr=image_bgr,
flatten=True)
X = pd.DataFrame(features)[feature_names]
y = model.predict(X)
segments = y.reshape(shape)
segments_bgr = [class2bgr(idx) for idx in segments.flatten()]
segments_bgr = np.array(segments_bgr).reshape(*shape, 3)
path = os.path.join(output_path, frame_folder)
if not os.path.exists(path):
os.makedirs(path)
image_and_segments_bgr = np.concatenate([image_bgr, segments_bgr],
axis=1)
# cv.imwrite(filename=os.path.join(path, image_name),
# img=image_bgr)
segments_filename = "camera" + camera_name + "_segments" + ".png"
cv.imwrite(
filename=os.path.join(path, segments_filename),
img=image_and_segments_bgr,
)
def load_model_and_generate_evaluation_images(
model_filename,
input_path: pathlib.Path,
output_path: pathlib.Path,
feature_names,
):
model = XGBClassifier()
model.load_model(model_filename)
for frame_folder in sorted(get_subdirectories(input_path)):
segment_names = [
f.name for f in frame_folder.iterdir()
if f.is_file() and f.name[1].isdigit()
]
if len(segment_names) != 0:
continue
for camera_name in ["60", "180", "300"]:
image_name = "camera" + camera_name + ".png"
print(frame_folder / image_name)
image_bgr = cv.imread(str(frame_folder / image_name))
features, shape = create_features(image_bgr=image_bgr,
flatten=True)
X = pd.DataFrame(features)[feature_names]
y = model.predict(X)
segments = y.reshape(shape)
segments_bgr = [class2bgr(idx) for idx in segments.flatten()]
segments_bgr = np.array(segments_bgr).reshape(*shape, 3)
path = output_path / frame_folder.name
if not path.exists():
path.mkdir(parents=True)
image_and_segments_bgr = np.concatenate([image_bgr, segments_bgr],
axis=1)
segments_filename = "camera" + camera_name + "_segments" + ".png"
cv.imwrite(
filename=str(path / segments_filename),
img=image_and_segments_bgr,
)
def predict(inputs):
main_category, category, goal, country, currency, today = inputs
""" Encode inputs """
encoder = load(ENCODER_PATH)
inputs2enc = np.array([category, main_category, currency, country]).reshape(1, -1)
inputs_encoded = encoder.transform(inputs2enc)
""" Stack """
numericals = np.array([goal, today.day, today.month]).reshape(1, -1)
final_inputs = np.hstack([numericals, inputs_encoded]).astype(np.float32)
""" Load model and predict """
model = XGBClassifier(seed=42)
model.load_model(MODEL_PATH)
result = model.predict_proba(final_inputs)
return result
def predict(age,
count,
diagnosis,
fpath='static/model/HFEA_model_{}',
nmodels=5):
"""
Loads and predicts from the models.
Parameters:
-----------
age : int,
Age of the patient in years.
count : int,
The number of Oocytes (eggs) collected following the treatment.
diagnosis : str,
The patients infertility diagnosis, must be one of Ovulatory disorder,
Male factor, Endometriosis or Unexplained.
fpath : str (default='static/model/HFEA_model_{}'),
Path to the models.
nmodel : int (default=5),
The number of models (i.e., the number of folds used in the
cross-validation proceedure during training).
Returns:
--------
pred, float:
"""
age_group = map_age_to_age_group(age)
# add the four infertility diagnosis features
infertility = create_infertility_feature(diagnosis)
X = np.r_[[age_group, count], infertility]
pred = 0
for i in range(5):
clf = XGBClassifier()
clf.load_model(f'static/model/HFEA_model_{i}')
pred += clf.predict_proba(X.reshape((1,-1)))[:,1][0] / nmodels
return pred
def predict(name, match_analysis_num, api_key):
#모델 및 데이터 로드
ss = joblib.load("model/standard_scaler.pkl")
xgb = XGBClassifier()
xgb.load_model("model/LOL_predict_xgb.bst")
match_df, player_stat, game_minute, win_lable = datapipe.collect_predict_data_by_name(
name, match_analysis_num, api_key)
del [[player_stat]]
gc.collect
if (win_lable[0] == -1):
return -1, -1
elif (win_lable[0] == -404):
return -404, -404
else:
#승률예측
match_scaled = ss.fit_transform(match_df)
win_rate = xgb.predict_proba(match_scaled)
real_win_rate = win_lable.mean()
predict_win_rate = win_rate[:, 1].mean()
return real_win_rate, predict_win_rate
def predict_xgb(data):
"""Perform prediction using trained model."""
model = XGBClassifier()
data = normalizator(prepare_data(data))
try:
model.load_model(
"cotopaxi/identification_models/proto_XGB_20201112.model")
except ValueError as exc:
raise CotopaxiException from exc(
"[!] Cannot load machine learning classifier!"
" This may be caused by incompatible version of tensorflow"
" (please install tensorflow version 2.2.0)!")
result = model.predict(data)
unique, counts = numpy.unique(result, return_counts=True)
devices = list()
for unit in unique:
devices.append(unit)
result_dict = dict(zip(devices, counts))
result_dict = sorted(result_dict.items(), key=lambda x: x[1], reverse=True)
result_class = result_dict[0][0]
return result_class, result_dict, counts.sum()
def run(self):
model_folder_path = '../model'
model_path = os.path.join(model_folder_path, 'xgb_final.pkl')
trained_model = XGBClassifier()
trained_model.load_model(model_path)
# read the processed features file
df_test = pd.read_csv(self.input().path)
# predict churn
prediction = trained_model.predict(df_test)
# putting prediction in dataframe as well as index ids (this is the
# the submission format)
submission = pd.DataFrame(data=prediction, columns=['churn'])
submission['churn'] = submission['churn'].map({1: 'yes', 0: 'no'})
submission.reset_index(inplace=True)
submission.rename(columns={'index': 'id'}, inplace=True)
submission['id'] = submission['id'] + 1
# write submission to file
submission.to_csv(self.output().path, index=False)
def stroke_predict(gender, age, hypertension, heart_disease, ever_married,
work_type, Residence_type, avg_glucose_level, bmi,
smoking_status):
# creating the pandas dataframe and replicating previous steps
a_dict = {
'gender': [int(gender)],
'age': [int(age)],
'hypertension': [int(hypertension)],
'heart_disease': [int(heart_disease)],
'ever_married': [int(ever_married)],
'work_type': [int(work_type)],
'Residence_type': [int(Residence_type)],
'avg_glucose_level': [float(avg_glucose_level)],
'bmi': [float(bmi)],
'smoking_status': [int(smoking_status)]
}
data = pd.DataFrame(a_dict)
data['gender'] = 1 if gender == 'male' else 0
data['ever_married'] = 1 if ever_married == 'Yes' else 0
work_mapping = {
'Self_employed': 3,
'Private': 2,
'children': 1,
'Govt_job': 0
}
data['work_type'] = data['work_type'].map(work_mapping)
data['Residence_type'] = 1 if Residence_type == 'Urban' else 0
smoke_mapping = {
'Unknown': 0,
'formerly smoked': 1,
'never_smoked': 2,
'smokes': 3
}
data['smoking_status'] = data['smoking_status'].map(smoke_mapping)
# after data has been replicated
xgb = XGBClassifier()
xgb.load_model("weights/stroke.model")
return xgb.predict(data)[0]
def start_pre(val_img_list, val_tar_list, type_class=minor_type_class):
real_class_pair_list = cut_class_pair
model_base_path = 'outs/'
result_list = [list() for i in range(len(val_img_list))]
config = Config()
result_max_item_list = [(0, 0) for i in range(len(val_img_list))]
for ci, class_pair in enumerate(type_class):
model_path = model_base_path + 'xgboost_model_per_class' + str(
class_pair) + '.pkl'
print('part ', ci, ' of ', len(real_class_pair_list))
clr = XGBClassifier()
clr.load_model(model_path)
y_p_x = clr.predict_proba(val_img_list)
pre_for_f1 = []
t_for_f1 = []
for i_ys, ys in enumerate(y_p_x):
if len(val_tar_list) > 0:
tail = ''
mid = ''
if class_pair in val_tar_list[i_ys]:
tail = '-----------'
if ys[1] >= 0.5:
mid = '||||||||'
print('ci ', ci, ' i_ys ', i_ys, ' pre ', ys, mid, ' c ',
class_pair, ' t ', val_tar_list[i_ys], tail)
else:
print('ci ', ci, ' i_ys ', i_ys, ' pre ', ys, ' c ',
class_pair)
sub_result = result_list[i_ys]
if ys[1] >= 0.5:
sub_result.append(class_pair)
pre_for_f1.append(1)
else:
pre_for_f1.append(0)
result_list[i_ys] = sub_result
max_item_idx, max_item_f = result_max_item_list[i_ys]
if ys[1] > max_item_f:
result_max_item_list[i_ys] = (class_pair, ys[1])
if len(val_tar_list) > 0:
for tar in val_tar_list:
if class_pair in tar:
t_for_f1.append(1)
else:
t_for_f1.append(0)
print('c ', class_pair, '---------f1 ',
f1_score(t_for_f1, pre_for_f1, average="macro"))
# print('sub ', ci, ' r:', sub_result)
pre_list = []
for this_sub_i, sub_result in enumerate(result_list):
print('this_sub_i ', this_sub_i, ' sub_result ', sub_result)
result_i = np.zeros(28)
for i_s, s in enumerate(sub_result):
result_i[s] += 1
# print('result_i ', result_i)
result = []
for i, r_i in enumerate(result_i):
if r_i == 1 and (i in type_class):
# print('i ', i, ' r_i ', r_i)
result.append(i)
if len(val_tar_list) > 0:
print('pre ', result, ' t ', val_tar_list[this_sub_i])
pre_list.append(result)
return pre_list, result_max_item_list
class FeatureClassifier():
def __init__(self, modelfile='featureclassifier.json'):
pwd = os.path.dirname(__file__)
self.model = XGBClassifier()
self.model.load_model(pwd + '/models/'+modelfile)
def check_errors(self, sim):
if sim.N_real < 4:
raise AttributeError("SPOCK Error: SPOCK only applicable to systems with 3 or more planets")
def predict_stable(self, sim, n_jobs=-1):
"""
Predict whether passed simulation will be stable over 10^9 orbits of the innermost planet.
Parameters:
sim (rebound.Simulation): Orbital configuration to test
n_jobs (int): Number of cores to use for calculation (only if passing more than one simulation). Default: Use all available cores.
Returns:
float: Estimated probability of stability. Will return exactly zero if configuration goes
unstable within first 10^4 orbits.
"""
res = self.generate_features(sim, n_jobs=n_jobs)
try:
stable = np.array([r[1] for r in res])
features = [r[0] for r in res]
Nsims = len(sim)
except:
stable = np.array([res[1]])
features = [res[0]]
Nsims = 1
# We take the negligible hit of evaluating XGBoost for all systems, and overwrite prob=0 for ones that went unstable in the short integration at the end
# array of Ntrios x 10 features to evaluate with XGboost (Nsims*Ntriospersim x 10 features)
featurevals = np.array([[val for val in trio.values()] for system in features for trio in system])
probs = self.model.predict_proba(featurevals)[:,1] # take 2nd column for probability it belongs to stable class
# XGBoost evaluated a flattened list of all trios, reshape so that trios in same sim grouped
trios_per_sim = int(len(probs)/Nsims)
probs = probs.reshape((Nsims, trios_per_sim))
# Take the minimum probability of stability within the trios for each simulation
probs = np.min(probs, axis=1)
# Set probabilities for systems that went unstable within short integration to exactly zero
probs[~stable] = 0
if Nsims == 1:
return probs[0]
else:
return probs
def generate_features(self, sim, n_jobs=-1):
"""
Generates the set of summary features used by the feature classifier for prediction.
Parameters:
sim (rebound.Simulation): Orbital configuration to test
n_jobs (int): Number of cores to use for calculation (only if passing more than one simulation). Default: Use all available cores.
Returns:
List of OrderedDicts: A list of sets of features for each adjacent trio of planets in system.
Each set of features is an ordered dictionary of 10 summary features. See paper.
stable (int): An integer for whether the N-body integration survived the 10^4 orbits (1) or
went unstable (0).
"""
if isinstance(sim, rebound.Simulation):
sim = [sim]
args = []
if len(set([s.N_real for s in sim])) != 1:
raise ValueError("If running over many sims at once, they must have the same number of particles!")
for s in sim:
s = s.copy()
init_sim_parameters(s)
minP = np.min([p.P for p in s.particles[1:s.N_real]])
self.check_errors(s)
trios = [[j,j+1,j+2] for j in range(1,s.N_real-2)] # list of adjacent trios
featureargs = [10000, 80, trios]
args.append([s, featureargs])
def run(params):
sim, featureargs = params
triofeatures, stable = features(sim, featureargs)
return triofeatures, stable
if len(args) == 1: # single sim
res = run(args[0]) # stable will be 0 if an orbit is hyperbolic
else:
if n_jobs == -1:
n_jobs = cpu_count()
#pool = ThreadPool(n_jobs)
res = map(run, args)
return list(res)
'model__colsample_bylevel': (0.01, 1.0, 'uniform'),
'model__learning_rate': (0.01, 1.0, 'log-uniform'),
'model__n_estimators': Integer(60, 400),
'model__max_depth': Integer(3, 12),
# 'model__scale_pos_weight': Real(1, 1000, 'log-uniform'), only binary
'model__min_child_weight': Integer(1, 15),
'model__gamma': Real(0.1, 3),
'model__alpha': Real(0, 1),
'model__lambda': Real(0, 1),
'model__subsample': Real(0.3, 1),
'model__colsample_bytree': Real(0, 1),
'model__colsample_bynode': Real(0, 1)
}
xg = XGBClassifier()
xg.load_model('XGBoost_model.json')
xgb_search_prev = {
'model': [xg],
# 'model__learning_rate': (0.01, 1.0, 'log-uniform'),
# 'model__min_child_weight': (0, 10),
# 'model__max_delta_step': Integer(0, 20),
# 'model__colsample_bytree': (0.01, 1.0, 'uniform'),
# 'model__colsample_bylevel': (0.01, 1.0, 'uniform'),
# 'model__n_estimators': Integer(100, 200),
# 'model__scale_pos_weight': Real(1, 1000, 'log-uniform'),
# 'model__min_child_weight': Integer(1, 10),
# 'model__gamma': Integer(1, 5),
# 'model__subsample': Real(0.3, 1),
# 'model__colsample_bytree': Real(0.1, 1),
# 'model__max_depth': Integer(6, 12)
RED = '\u001b[31m'
GREEN = '\u001b[32m'
BLUE = '\u001b[34m'
RESET = '\033[0m'
xgb_mod = XGBClassifier(booster='dart',
tree_method="gpu_hist",
n_estimators=300,
learning_rate=0.05,
predictor='gpu_predictor',
eval_metric='logloss',
max_depth=3,
gpu_id=0)
xgb_mod.load_model('CVD_mod')
cvd_df = pd.read_csv('cardio_train.csv', sep=';', index_col=0)
cvd_df['age'] = cvd_df['age'] / 365.24
cvd_df['gender'] = cvd_df['gender'] - 1
cvd_df = cvd_df[(cvd_df['ap_lo'] <= 370) & (cvd_df['ap_lo'] > 0)]
cvd_df = cvd_df[(cvd_df['ap_hi'] <= 370) & (cvd_df['ap_hi'] > 0)]
cvd_df = cvd_df[cvd_df['ap_hi'] >= cvd_df['ap_lo']]
cvd_df.reset_index(drop=True, inplace=True)
X_train, X_test, y_train, y_test = train_test_split(cvd_df.drop(['cardio'],
axis=1),
cvd_df['cardio'],
class XGBoost(BaseAlgorithm):
def __init__(self, algorithm_settings, problem_type):
super().__init__(algorithm_settings)
self.problem_type = problem_type
def build(self):
if self.problem_type == SupervisedTask.regression:
self.build_regression_model()
elif self.problem_type == SupervisedTask.classification:
self.build_classification_model()
else:
raise TypeError('Unknown problem_type')
def build_regression_model(self):
from xgboost import XGBRegressor
self.model = XGBRegressor(
max_depth=self.algorithm_settings.max_depth,
learning_rate=self.algorithm_settings.learning_rate,
n_estimators=self.algorithm_settings.n_estimators,
objective=self.algorithm_settings.objective,
booster=self.algorithm_settings.booster,
n_jobs=self.algorithm_settings.n_jobs,
gamma=self.algorithm_settings.gamma,
min_child_weight=self.algorithm_settings.min_child_weight,
max_delta_step=self.algorithm_settings.max_delta_step,
subsample=self.algorithm_settings.subsample,
reg_alpha=self.algorithm_settings.reg_alpha,
reg_lambda=self.algorithm_settings.reg_lambda,
random_state=self.algorithm_settings.random_state)
def build_classification_model(self):
from xgboost import XGBClassifier
self.model = XGBClassifier(
max_depth=self.algorithm_settings.max_depth,
learning_rate=self.algorithm_settings.learning_rate,
n_estimators=self.algorithm_settings.n_estimators,
objective=self.algorithm_settings.objective,
booster=self.algorithm_settings.booster,
n_jobs=self.algorithm_settings.n_jobs,
gamma=self.algorithm_settings.gamma,
min_child_weight=self.algorithm_settings.min_child_weight,
max_delta_step=self.algorithm_settings.max_delta_step,
subsample=self.algorithm_settings.subsample,
reg_alpha=self.algorithm_settings.reg_alpha,
reg_lambda=self.algorithm_settings.reg_lambda,
random_state=self.algorithm_settings.random_state)
def train(self, train_x, train_y, settings):
self.model.fit(train_x,
train_y,
eval_metric=self.algorithm_settings.eval_metric)
self.save(settings)
def evaluate(self, test_x):
prediction = self.model.predict(test_x)
prediction = prediction.reshape(-1, 1)
return prediction
def load(self, model_path):
self.model.load_model(fname=model_path)
def save(self, settings):
model_save_dir = os.path.join(settings.models_path, 'xgboost_models')
os.makedirs(model_save_dir, exist_ok=True)
model_name = self.get_model_name(settings)
save_path = os.path.join(model_save_dir, model_name)
self.model.save_model(fname=save_path)
print(f"Model saved to: {save_path}")
def get_model_name(self, settings):
if settings.problem_type == SupervisedTask.regression:
return 'regression_model.xgb'
else:
return 'classification_model.xgb'
def get_model():
xgboost_quora_model = BASE_URL + "/xgboost_xcfl_quora_model.model"
x_cfl = XGBClassifier()
x_cfl.load_model(xgboost_quora_model)
return x_cfl
# rmse, mae, logloss, error(설명 error가 accuracy), auc(설명 accuracy친구)
results = model.evals_result()
# print("eval's results : ", results)
# print("r2 Score : %.2f%%:" %(r2*100.0))
y_pred = model.predict(x_test)
acc = accuracy_score(y_pred, y_test)
print("acc : ", acc)
#####################################################################################################
# import pickle # 파이썬에서 제공한다.
# from joblib import dump, load
# import joblib
# pickle.dump(model, open("./model/xgb_save/cancer.pickle.dat", "wb")) # wb형식으로 저장하겠다.
# joblib.dump(model, "./model/xgb_save/cancer.joblib.dat")
model.save_model("./model/xgb_save/cancer.xgb.model")
print("저장됬다.")
# model2 = pickle.load(open("./model/xgb_save/cancer.pickle.dat", "rb"))
# model2 = joblib.load("./model/xgb_save/cancer.joblib.dat")
model2 = XGBClassifier()
model2.load_model("./model/xgb_save/cancer.xgb.model")
print('불러왔다.')
y_pred = model2.predict(x_test)
acc = accuracy_score(y_pred, y_test)
print("acc : ", acc)
provides recommendations for portfolio action, backtests, and incrementally trains
the given XGBoost Classifier Model'''
from xgboost import XGBClassifier
import cpdb
import pandas as pd
import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import Imputer
from sklearn.metrics import zero_one_loss
from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV
btc_model = XGBClassifier()
eth_model = XGBClassifier()
btc_model = btc_model.load_model('btc_model.bin')
eth_model = eth_model.load_model('eth_model.bin')
model = {
'btc': btc_model,
'eth': eth_model,
}
'''Delivers a recommendation based on model classification'''
def get_recommendation(coinName, features):
coin_model = model.get(coinName)
recommendation = coin_model.predict(features)
return recommendation
def is_spam(data, mode=2, classifier='manual'):
if (classifier == 'manual'):
message_body = data
if (mode != 2):
message_body = get_email(data, mode)
clean_message = clean_message_no_html(message_body,
stop_words=set(
stopwords.words('english')))
word_columns_df = pd.DataFrame.from_records([clean_message])
word_columns_df.index.name = 'DOC_ID'
word_index = pd.Index(vocab.VOCAB_WORD)
sparse_matrix = make_sparse_matrix(word_columns_df,
word_index).groupby([
'DOC_ID', 'WORD_ID'
]).sum().reset_index().to_numpy()
full_matrix = make_full_matrix(sparse_matrix,
vocab.shape[0]).to_numpy()
spam_email_prob = PROB_SPAM
ham_email_prob = 1 - PROB_SPAM
# denominator = 1
for j in range(full_matrix.shape[1]):
if full_matrix[0, j] > 0:
if prob_token_spam[j] > 0:
spam_email_prob = spam_email_prob * \
(prob_token_spam[j]**full_matrix[0, j])
if spam_email_prob == 0:
spam_email_prob = prev_spam
ham_email_prob = prev_ham
break
if prob_token_ham[j] > 0:
ham_email_prob = ham_email_prob * \
(prob_token_ham[j]**full_matrix[0, j])
if ham_email_prob == 0:
spam_email_prob = prev_spam
ham_email_prob = prev_ham
break
prev_spam = spam_email_prob
prev_ham = ham_email_prob
# denominator = denominator * prob_all_tokens[j]
# print(spam_email_prob/denominator > ham_email_prob/denominator)
print(spam_email_prob > ham_email_prob)
# joint_log_spam = full_matrix.dot(
# np.log(prob_token_spam+0.000000000000001) - np.log(prob_all_tokens+0.000000000000001)) + np.log(PROB_SPAM)
# print(joint_log_spam)
# joint_log_ham = full_matrix.dot(
# np.log(prob_token_ham+0.000000000000001) - np.log(prob_all_tokens+0.000000000000001)) + np.log(1 - PROB_SPAM)
# print(joint_log_ham)
elif (classifier == 'xgb'):
xgb_classifier = XGBClassifier()
xgb_classifier.load_model('./XGB.model')
data_list = []
data_list.append(data)
doc_term_matrix = vectorizer.transform(data_list)
print(xgb_classifier.predict(doc_term_matrix)[0] == 1)
class Classifier:
# for initializing train and test sets, classifier and accuracy score
# Change method to gpu_hist if you want xgboost to run on a GPU
def __init__(self,
params={
'objective': 'reg:squarederror',
'verbosity': 0
}):
self.X_train = []
self.X_labels = []
self.test = []
self.test_labels = []
self.model = XGBClassifier(**params)
self.prediction = 0
self.error = 0
def size(self):
if isinstance(self.X_train, np.ndarray):
return self.X_train.size
return len(self.X_train)
# adding the data points
def input_train(self, features, feature):
if isinstance(self.X_train, np.ndarray) and self.X_train.size > 0:
self.X_train = self.X_train.tolist()
self.X_labels = self.X_labels.tolist()
self.X_train.append(features)
self.X_labels.append(feature)
# train the data
def train(self):
self.X_train = np.asarray(self.X_train)
self.X_labels = np.asarray(self.X_labels)
self.model.fit(self.X_train, self.X_labels)
def train_eval(self, metric='error'):
self.X_train = np.asarray(self.X_train)
self.X_labels = np.asarray(self.X_labels)
X_train, X_test, y_train, y_test = train_test_split(self.X_train,
self.X_labels,
test_size=0.33)
self.model.fit(X_train,
y_train,
eval_set=[(X_train, y_train), (X_test, y_test)],
eval_metric=metric)
evals_result = self.model.evals_result()
if metric == 'error':
validations = []
for val in evals_result.values():
lst = val.get("error")
validations.append(sum(lst) / len(lst))
return 1 - (sum(validations) / len(validations))
else:
validations = []
for val in evals_result.values():
lst = val.get(metric)
validations.append(lst[-1])
return validations
# input test labels if you want to check accuracy
def label(self, label):
self.test_labels.append(label)
def input_test(self, features):
if isinstance(self.test, np.ndarray) and self.test.size > 0:
self.test = self.test.tolist()
self.test.append(features)
# test data
def predict(self):
if not isinstance(self.test, np.ndarray):
self.test = np.asarray(self.test)
self.prediction = self.model.predict(self.test)
return self.prediction
def predict_proba(self):
if not isinstance(self.test, np.ndarray):
self.test = np.asarray(self.test)
self.prediction = self.model.predict_proba(self.test)
return self.prediction
# if you have the test labels you can check the error rate (you want error close to 0)
def check_error(self):
self.test_labels = np.asarray(self.test_labels)
self.error = metrics.mean_absolute_error(self.test_labels,
self.prediction)
return self.error
# save classifier
def save_classifier(self, file):
self.model.save_model(file)
# open saved classifier
def open_classifier(self, file):
self.model.load_model(file)
# removes all training data
def clean_train(self):
self.X_train = []
self.X_labels = []
# removes all testing data
def clean_test(self):
self.test = []
self.test_labels = []
def predict_sent(vecs, xgb_model_analyze):
xgb = XGBC()
xgb.load_model(xgb_model_analyze)
pred = predict(vecs, w2v_model, xgb, 300)
df = pd.DataFrame(pred, columns=['sent'])
return df
