def test_fetch_error():
with raises(ValueError, match='is not a dataset available.'):
fetch_datasets(filter_data=tuple(['rnd']))
with raises(ValueError, match='dataset with the ID='):
fetch_datasets(filter_data=tuple([-1]))
with raises(ValueError, match='dataset with the ID='):
fetch_datasets(filter_data=tuple([100]))
with raises(ValueError, match='value in the tuple'):
fetch_datasets(filter_data=tuple([1.00]))
def test_fetch_error():
with raises(ValueError, match='is not a dataset available.'):
fetch_datasets(filter_data=tuple(['rnd']))
with raises(ValueError, match='dataset with the ID='):
fetch_datasets(filter_data=tuple([-1]))
with raises(ValueError, match='dataset with the ID='):
fetch_datasets(filter_data=tuple([100]))
with raises(ValueError, match='value in the tuple'):
fetch_datasets(filter_data=tuple([1.00]))
def __extract_binarized_imbalanced_datasets():
for dataset_name, dataset_values in fetch_datasets().items():
write_dataset_to_csv("./binarized-datasets/" + dataset_name + ".csv", dataset_values)
# if __name__ == '__main__':
# __labelize_dataset("E:/python-workspace/resampler/binarized-datasets/"
# "2_Class_Data_February_Cleaned_with_custom_header.csv",
# "E:/python-workspace/resampler/binarized-datasets/custom_ds.csv")
#__extract_binarized_imbalanced_datasets()
def load_dataset(data_name):
load_data = fetch_datasets(verbose=True)[data_name]
print(load_data.data.shape)
print(Counter(load_data.target))
X = pd.DataFrame(load_data.data)
y = pd.DataFrame(load_data.target, columns=['Label'])
return X, y
def load_datasets(dataset, names):
if dataset == "wilt":
X, y, cl_names = load_wilt()
elif dataset == "adult":
X, y, cl_names = load_adult()
elif dataset == "diabetes":
X, y, cl_names = load_diabetes()
elif dataset == "phoneme":
X, y, cl_names = load_phoneme()
elif dataset == "mushroom":
X, y, cl_names = load_mushroom()
elif dataset == "electricity":
X, y, cl_names = load_electricity()
elif dataset == "speeddating":
X, y, cl_names = load_speed_dating()
elif dataset == "credit":
X, y, cl_names = load_credit()
elif dataset == "eeg_eye":
X, y, cl_names = load_eeg_eye()
elif dataset == "spam":
X, y, cl_names = load_spam()
elif dataset == "skin":
X, y, cl_names = load_skin()
elif dataset == "bank":
X, y, cl_names = load_bank()
elif dataset == "kdd":
X, y, cl_names = load_kdd()
elif dataset == "landsatM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "musk2":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "spliceM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "semeion_orig":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "waveformM":
X, y, cl_names = load_mat_data(dataset)
else:
from imblearn import datasets
data = datasets.fetch_datasets()[dataset]
cl_names = [
"feature_" + str(i) for i in range(0, data['data'].shape[1])
]
X = data['data']
y = data['target']
y[y != 1] = 0
names.add(dataset)
output = []
output.append(X.shape[0])
output.append(X.shape[1])
output.append(float(format(len(abs(y[y != 1])) / sum(y[y == 1]), '.2f')))
return output
def get_dataset(dataset):
if dataset == "wilt":
X, y, cl_names = load_wilt()
elif dataset == "adult":
X, y, cl_names = load_adult()
elif dataset == "diabetes":
X, y, cl_names = load_diabetes()
elif dataset == "phoneme":
X, y, cl_names = load_phoneme()
elif dataset == "mushroom":
X, y, cl_names = load_mushroom()
elif dataset == "electricity":
X, y, cl_names = load_electricity()
elif dataset == "speeddating":
X, y, cl_names = load_speed_dating()
elif dataset == "credit":
X, y, cl_names = load_credit()
elif dataset == "eeg_eye":
X, y, cl_names = load_eeg_eye()
elif dataset == "spam":
X, y, cl_names = load_spam()
elif dataset == "skin":
X, y, cl_names = load_skin()
elif dataset == "bank":
X, y, cl_names = load_bank()
elif dataset == "kdd":
X, y, cl_names = load_kdd()
elif dataset == "landsatM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "musk2":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "spliceM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "semeion_orig":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "waveformM":
X, y, cl_names = load_mat_data(dataset)
elif dataset not in ['bloob', 'circle', 'moon']:
from imblearn import datasets
data = datasets.fetch_datasets()[dataset]
cl_names = [
"feature_" + str(i) for i in range(0, data['data'].shape[1])
]
X = data['data']
y = data['target']
y[y != 1] = 0
return X, y, cl_names
def test_documentation_example():
"""Test basic code example shown in documentation"""
from imblearn.datasets import fetch_datasets
datasets = fetch_datasets(filter_data=['oil'])
X, y = datasets['oil']['data'], datasets['oil']['target']
labels, counts = np.unique(y, return_counts=True)
assert counts[0] > counts[1]
kmeans_smote = KMeansSMOTE(kmeans_args={'n_clusters': 100},
smote_args={'k_neighbors': 10})
X_resampled, y_resampled = kmeans_smote.fit_sample(X, y)
labels, counts = np.unique(y_resampled, return_counts=True)
assert counts[0] == counts[1]
def print_examples():
ts = fetch_datasets()['thyroid_sick']
print(ts.data.shape)
target_classes = sorted(Counter(ts.target).items())
print(sorted(Counter(ts.target).items()))
ds = load_ds('../datasets/binarized-datasets/thyroid_sic_2.data')
labels = ['Target classes']
healty, sick = ([len(list(filter(lambda x: x[-1] == 0, ds)))],
[len(list(filter(lambda x: x[-1] == 1, ds)))])
# healty = [target_classes[0][1]]
# sick = [target_classes[1][1]]
x = np.arange(len(labels)) # the label locations
width = 0.35 # the width of the bars
fig, ax = plt.subplots()
rects1 = ax.bar(x - width / 2, healty, width, label='Healthy')
rects2 = ax.bar(x + width / 2, sick, width, label='Sick')
# Add some text for labels, title and custom x-axis tick labels, etc.
ax.set_ylabel('Number of samples')
ax.set_xticks(x)
ax.set_xticklabels(labels)
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.xaxis.set_ticks_position('bottom')
ax.yaxis.set_ticks_position('left')
ax.legend()
def autolabel(rects):
"""Attach a text label above each bar in *rects*, displaying its height."""
for rect in rects:
height = rect.get_height()
ax.annotate(
'{}'.format(height),
xy=(rect.get_x() + rect.get_width() / 2, height),
xytext=(0, 3), # 3 points vertical offset
textcoords="offset points",
ha='center',
va='bottom')
autolabel(rects1)
autolabel(rects2)
fig.tight_layout()
plt.show()
def trial(name, sampling_strategy, k_neighbors, n_jobs):
setup = f'''
from imblearn.datasets import fetch_datasets
from imblearn.over_sampling import SMOTE
sampling_strategy = '{sampling_strategy}'
k_neighbors = {k_neighbors}
n_jobs = {n_jobs}
dataset = fetch_datasets()['{name}']
X, y = dataset.data, dataset.target
smote = SMOTE(sampling_strategy, k_neighbors=k_neighbors, n_jobs=n_jobs, random_state=0)
'''
setup = textwrap.dedent(setup).strip()
t = timeit('smote.fit_resample(X, y)', setup=setup, number=100)
dataset = fetch_datasets()[name]
X, y = dataset.data, dataset.target
smote = SMOTE(sampling_strategy, k_neighbors=k_neighbors, n_jobs=n_jobs, random_state=0)
X_resampled, y_resampled = smote.fit_resample(X, y)
idx = -len(X)
X_new, y_new = X_resampled[idx:], y_resampled[idx:]
return X_new, y_new, t
from sklearn.metrics import balanced_accuracy_score
from imblearn.metrics import geometric_mean_score
from matplotlib import pyplot as plt
import seaborn as sns
from seaborn import scatterplot
from numpy import where
from collections import Counter
import numpy as np
# get_ipython().run_line_magic('matplotlib', 'inline')
st.title("Skripsiku")
name = 'pen_digits'
dataset = fetch_datasets()[name]
X = dataset.data
y = dataset.target
df = pd.concat([pd.DataFrame(X), pd.DataFrame(y)], axis=1)
st.subheader('Dataset Name :')
st.write(name)
st.write(df)
cv = StratifiedKFold(n_splits=5, random_state=1, shuffle=True)
for train_index, test_index, in cv.split(X, y):
# st.write("Train: \n", train_index, "\nValidation:\n", test_index)
X_train, X_test = X[train_index], X[test_index]
def fetch(*args, **kwargs):
return fetch_datasets(*args, download_if_missing=True, **kwargs)
def templet(sampler_name, sample_ratio):
"""
模板方法
:param sampler_name: 采样算法名
:param sample_ratio: 采样比例
:return:
"""
dataset = fetch_datasets()['satimage']
X = dataset.data
y = dataset.target
# 起始时间
start_time = time.time()
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
sb = None
if sampler_name == 'CART':
sb = DummySampler()
elif sampler_name == 'SMOTE':
sb = SMOTE(N=sample_ratio, k_neighbors=5, random_state=42)
elif sampler_name == 'Border1':
sb = BorderSMOTE(N=sample_ratio,
m_neighbors=9,
k_neighbors=5,
random_state=42,
kind='borderline1')
elif sampler_name == 'Border2':
sb = BorderSMOTE(N=sample_ratio,
m_neighbors=9,
k_neighbors=5,
random_state=42,
kind='borderline2')
elif sampler_name == 'ADASYN':
sb = ADASYN(bata=sample_ratio, k_neighbors=5, random_state=42)
elif sampler_name == 'Safe-level':
sb = SafeLevelSMOTE(N=sample_ratio, k_neighbors=5, random_state=42)
else:
pass
X_res, y_res = sb.fit_sample(X_train_minmax, y[train]) # 采样
model = tree.DecisionTreeClassifier(max_depth=8,
min_samples_split=10,
random_state=42)
model.fit(X_res, y_res)
predict = model.predict(X_test_minmax)
probability = model.predict_proba(X_test_minmax)[:, 1]
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
# write2dic
fill_dic('precision', sampler_name, sample_ratio, precision)
fill_dic('recall', sampler_name, sample_ratio, recall)
fill_dic('f1', sampler_name, sample_ratio, f1)
fill_dic('auc', sampler_name, sample_ratio, auc)
fill_dic('gmean', sampler_name, sample_ratio, gmean)
print('%s %.1f building id transforming took %fs!' %
(sampler_name, sample_ratio, time.time() - start_time))
ax.text(j, i, format(cm[i, j], fmt),
horizontalalignment="center",
color="white" if cm[i, j] > thresh else "black")
ax.set_ylabel('True label')
ax.set_xlabel('Predicted label')
###############################################################################
# Load an imbalanced dataset
###############################################################################
# We will load the UCI SatImage dataset which has an imbalanced ratio of 9.3:1
# (number of majority sample for a minority sample). The data are then split
# into training and testing.
satimage = fetch_datasets()['satimage']
X, y = satimage.data, satimage.target
X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y,
random_state=0)
###############################################################################
# Classification using a single decision tree
###############################################################################
# We train a decision tree classifier which will be used as a baseline for the
# rest of this example.
###############################################################################
# The results are reported in terms of balanced accuracy and geometric mean
# which are metrics widely used in the literature to validate model trained on
# imbalanced set.
print(f"Total train time: {clf_copy._total_train_time}")
print(f"Total fit time: {clf_copy._fit_time}")
print(f"Total iterations: {clf_copy._iter_count}")
print(f"Final Sampling Probability: {clf_copy._sampling_proba}")
# Save results
try:
results_df = pd.concat(
[pd.read_csv(output_path, index_col=0),
pd.DataFrame(results)])
except FileNotFoundError:
results_df = pd.DataFrame(results)
results_df.to_csv(output_path)
if __name__ == "__main__":
database = fetch_datasets()
for dataset_name in DATASETS:
for estimator_ in ESTIMATORS:
for cost_scaling_ in COST_SCALINGS:
try:
run_experiment(dataset=database[dataset_name],
estimator=estimator_,
cost_scaling=cost_scaling_,
output_path=OUTPUT_PATH,
verbose=True)
print(
f"Completed experiment on {dataset_name} dataset with {type(estimator_)} model "
+ f"and cost scaling {cost_scaling_}")
except BlockingIOError:
print(
f"Error running experiment on {dataset_name} dataset with {type(estimator_)} model "
cut_perc = 0.1
cut_intervals = (512, 256, 128, 64)
continuation = True
# Nia to be used in the experiment
evos = [
GreyWolfOptimizer, SelfAdaptiveDifferentialEvolution, GeneticAlgorithm,
EvolutionStrategyMpL, ParticleSwarmAlgorithm
]
# Datasets
dataset_names = [
'libras_move', 'spectrometer', 'optical_digits', 'oil', 'ozone_level',
'arrhythmia', 'us_crime', 'yeast_ml8'
]
datasets = fetch_datasets(verbose=True)
with open('./results/results_all5.csv', 'w') as f:
print(
'Algorithm,Dataset,Fold,Accuracy,Fscore,TrainingTime,NoFeatures,Solution'
)
print(
'Algorithm,Dataset,Fold,Accuracy,Fscore,TrainingTime,NoFeatures,Solution',
file=f)
# For each dataset
for dataset_name in dataset_names:
dataset = datasets[dataset_name]
scaler = MinMaxScaler() # Scale it
dataset.data = scaler.fit_transform(dataset.data)
skf = StratifiedKFold(n_splits=10,
def model(boosting_name, data_name, classifier_name, cv_name, mode):
"""
模板方法
:param boosting_name: 集成学习的方法
:param data_name: 数据集名称
:param classifier_name: 使用的基分类器
:param cv_name: 交叉验证模式
:param mode: 采样模式
:return:
"""
# 加载数据
if data_name in fetch_datasets().keys():
dataset = fetch_datasets()[data_name]
X = dataset.data
y = dataset.target
print(Counter(y))
else:
# 加载自定义数据
df = pd.read_csv('../imbalanced_data/%s.csv' % data_name, header=None)
array = df.values.astype(float)
X = array[:, 0:array.shape[1] - 1]
y = array[:, -1]
print(Counter(y))
base = None
if classifier_name == 'CART':
base = tree.DecisionTreeClassifier(max_depth=8,
random_state=42,
min_samples_split=10)
elif classifier_name == 'svm':
base = svm.SVC()
else:
pass
# 起始时间
start_time = time.time()
cv = None
if cv_name == 'StratifiedKFold':
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
elif cv_name == 'RepeatedStratifiedKFold':
cv = RepeatedStratifiedKFold(n_repeats=10,
n_splits=10,
random_state=42)
else:
pass
mean_tpr = 0.0
mean_fpr = np.linspace(0, 1, 100) # 插值点(保证每一折的fpr和tpr相同)
aucs = []
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
classifier = None
if boosting_name == 'CART':
classifier = base
elif boosting_name == 'Bagging':
classifier = BaggingClassifier(base_estimator=base,
n_estimators=40)
elif boosting_name == 'BalancedBagging':
classifier = BalancedBaggingClassifier(base_estimator=base,
ratio='auto',
replacement=True,
random_state=42)
elif boosting_name == 'Adaboost':
classifier = AdaBoostClassifier(base_estimator=base,
n_estimators=40)
elif boosting_name == 'Random Forest':
classifier = RandomForestClassifier(max_depth=8,
min_samples_split=10,
n_estimators=40,
random_state=42)
elif boosting_name == 'EasyEnsemble':
model_under(boosting_name, X_train_minmax, y[train], X_test_minmax,
y[test])
continue
elif boosting_name == 'BalanceCascade':
model_under(boosting_name, X_train_minmax, y[train], X_test_minmax,
y[test])
continue
elif boosting_name == 'SMOTEBoost':
classifier = SMOTEBoost(rate=100,
n_estimators=40,
weak_estimator=base,
random_state=42,
class_dist=False)
elif boosting_name == 'RUSBoost':
classifier = RUSBoost(ratio=50,
n_estimators=40,
weak_estimator=base,
random_state=42,
class_dist=False)
else:
pass
classifier.fit(X_train_minmax, y[train]) # 采样
predict = classifier.predict(X_test_minmax)
probability = classifier.predict_proba(X_test_minmax)[:, 1]
# 指标计算
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
accuracy = metrics.accuracy_score(y[test], predict)
# -------------step6.计算每一折的ROC曲线和PR曲线上的点 -------------
fpr, tpr, thresholds = metrics.roc_curve(y[test], probability)
# 对mean_tpr在mean_fpr处进行插值，通过scipy包调用interp()函数
mean_tpr += interp(mean_fpr, fpr, tpr)
mean_tpr[0] = 0.0 # 为什么？
roc_auc = metrics.auc(fpr, tpr)
aucs.append(roc_auc)
# write2dic
fill_dic('precision', boosting_name, precision)
fill_dic('recall', boosting_name, recall)
fill_dic('f1', boosting_name, f1)
fill_dic('auc', boosting_name, auc)
fill_dic('gmean', boosting_name, gmean)
if boosting_name != 'EasyEnsemble' and boosting_name != 'BalanceCascade':
# 将frp和tpr写入文件
# 在mean_fpr100个点，每个点处插值插值多次取平均
mean_tpr /= cv.get_n_splits()
# 坐标最后一个点为（1,1）
mean_tpr[-1] = 1.0
# 计算平均AUC值
mean_auc = metrics.auc(mean_fpr, mean_tpr)
# 将平均fpr和tpr拼接起来存入文件
filename = './ROC/{data_name}/{mode}/{base_classifier}/{sampler}.csv'. \
format(data_name=data_name, mode=mode, base_classifier=classifier_name, sampler=boosting_name)
# 将文件路径分割出来
file_dir = os.path.split(filename)[0]
# 判断文件路径是否存在，如果不存在，则创建，此处是创建多级目录
if not os.path.isdir(file_dir):
os.makedirs(file_dir)
# # 然后再判断文件是否存在，如果不存在，则创建
# if not os.path.exists(filename):
# os.system(r'touch %s' % filename)
# 将结果拼合起来
all = np.c_[mean_fpr, mean_tpr]
np.savetxt(filename, all, delimiter=',', fmt='%f')
print('%s building id transforming took %fs!' %
(boosting_name, time.time() - start_time))
from imblearn.under_sampling import NearMiss
from imblearn.metrics import classification_report_imbalanced
from sklearn.metrics import precision_score, recall_score, f1_score, roc_auc_score, accuracy_score, classification_report
from sklearn.ensemble import RandomForestClassifier
import numpy as np
def print_results(headline, true_value, pred):
print(headline)
print("accuracy: {}").format(accuracy_score(true_value, pred))
print("precision: {}").format(precision_score(true_value, pred))
print("recall: {}").format(recall_score(true_value, pred))
print("f1: {}").format(f1_score(true_value, pred))
data = fetch_datasets()['wine_quality']
X_train, X_test, y_train, y_test = train_test_split(data['data'],
data['target'],
random_state=2)
#build normal model
pipeline = make_pipeline(RandomForestClassifier(random_state=42))
model = piepline.fit(X_train, y_test)
prediction = model.predict(X_test)
#build model with SMOTE
smote_pipeline = make_pipeline_imb(SMOTE(random_state=4),
RandomForestClassifier(random_state=42))
smote_model = smote_pipeline.fit(X_train, y_train)
smote_prediction = smote_model.predict(X_test)
def test():
dic = {'recall': {'CART': [], 'SMOTE': [], 'Border1': [], 'Border2': [], 'ADASYN': [], 'Safe-level': []},
'precision': {'CART': [], 'SMOTE': [], 'Border1': [], 'Border2': [], 'ADASYN': [], 'Safe-level': []},
'f1': {'CART': [], 'SMOTE': [], 'Border1': [], 'Border2': [], 'ADASYN': [], 'Safe-level': []},
'auc': {'CART': [], 'SMOTE': [], 'Border1': [], 'Border2': [], 'ADASYN': [], 'Safe-level': []},
'gmean': {'CART': [], 'SMOTE': [], 'Border1': [], 'Border2': [], 'ADASYN': [], 'Safe-level': []}}
results = prettytable.PrettyTable(["Classifier", "Precision", 'Recall', 'AUC', 'F-measure', 'G-mean'])
# 加载数据
dataset = fetch_datasets()['satimage']
X = dataset.data
y = dataset.target
print(Counter(y))
# 随机种子，保证每次实验结果相同
np.random.seed(42)
# -------------------------------------------CART----------------------------------------------------
# 起始时间
start_time = time.time()
# 交叉验证CART
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# initialize CART
cart = tree.DecisionTreeClassifier(max_depth=8, min_samples_split=10, random_state=42)
# 归一化
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 训练
cart.fit(X_train_minmax, y[train])
# 测试
predict = cart.predict(X_test_minmax)
probability = cart.predict_proba(X_test_minmax)
cart_auc = metrics.roc_auc_score(y[test], probability[:, 1])
cart_precision = metrics.precision_score(y[test], predict)
cart_recall = metrics.recall_score(y[test], predict)
if cart_precision == 0:
cart_f1 = 0
else:
cart_f1 = 2 * (cart_precision * cart_recall) / (cart_precision + cart_recall)
cart_gmean = geometric_mean_score(y[test], predict)
dic['precision']['CART'].append(cart_precision)
dic['recall']['CART'].append(cart_recall)
dic['f1']['CART'].append(cart_f1)
dic['auc']['CART'].append(cart_auc)
dic['gmean']['CART'].append(cart_gmean)
print('CART building id transforming took %fs!' % (time.time() - start_time))
# ---------------------------------------------------SMOTE----------------------------------------------------------
# 起始时间
start_time = time.time()
# 交叉验证
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=10, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# preprocess
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# initialize sampler
sb = SMOTE(N=100, k_neighbors=5, random_state=42)
# sampling
X_res, y_res = sb.fit_sample(X_train_minmax, y[train])
# initialize classifier
model = tree.DecisionTreeClassifier(max_depth=8, min_samples_split=10, random_state=42)
# model = svm.SVC(class_weight={1: 20})
model.fit(X_res, y_res)
predict = model.predict(X_test_minmax)
probability = model.predict_proba(X_test_minmax)[:, 1]
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
dic['precision']['SMOTE'].append(precision)
dic['recall']['SMOTE'].append(recall)
dic['f1']['SMOTE'].append(f1)
dic['auc']['SMOTE'].append(auc)
dic['gmean']['SMOTE'].append(gmean)
print('SMOTE building id transforming took %fs!' % (time.time() - start_time))
# ---------------------------------------------Borderline-SMOTE1----------------------------------------------------
# 起始时间
start_time = time.time()
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 初始化采样器
sb = BorderSMOTE(N=100, m_neighbors=30, k_neighbors=5, random_state=42, kind='borderline1')
# 采样
X_res, y_res = sb.fit_sample(X_train_minmax, y[train])
model = tree.DecisionTreeClassifier(max_depth=8, min_samples_split=10, random_state=42)
model.fit(X_res, y_res)
predict = model.predict(X_test_minmax)
probability = model.predict_proba(X_test_minmax)[:, 1]
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
dic['precision']['Border1'].append(precision)
dic['recall']['Border1'].append(recall)
dic['f1']['Border1'].append(f1)
dic['auc']['Border1'].append(auc)
dic['gmean']['Border1'].append(gmean)
print('BorderSmote1 building id transforming took %fs!' % (time.time() - start_time))
# ---------------------------------------------Borderline-SMOTE2----------------------------------------------------
# 起始时间
start_time = time.time()
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 初始化采样器
sb = BorderSMOTE(N=100, m_neighbors=30, k_neighbors=5, random_state=42, kind='borderline2')
# 采样
X_res, y_res = sb.fit_sample(X_train_minmax, y[train])
model = tree.DecisionTreeClassifier(max_depth=8, min_samples_split=10, random_state=42)
model.fit(X_res, y_res)
predict = model.predict(X_test_minmax)
probability = model.predict_proba(X_test_minmax)[:, 1]
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
dic['precision']['Border2'].append(precision)
dic['recall']['Border2'].append(recall)
dic['f1']['Border2'].append(f1)
dic['auc']['Border2'].append(auc)
dic['gmean']['Border2'].append(gmean)
print('BorderSmote2 building id transforming took %fs!' % (time.time() - start_time))
# ---------------------------------------------ADASYN---------------------------------------------------------------
# 起始时间
start_time = time.time()
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 训练
sb = ADASYN(bata=0.1, k_neighbors=5, random_state=42)
# 预测
X_res, y_res = sb.fit_sample(X_train_minmax, y[train])
model = tree.DecisionTreeClassifier(max_depth=8, min_samples_split=10, random_state=42)
model.fit(X_res, y_res)
predict = model.predict(X_test_minmax)
probability = model.predict_proba(X_test_minmax)[:, 1]
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
dic['precision']['ADASYN'].append(precision)
dic['recall']['ADASYN'].append(recall)
dic['f1']['ADASYN'].append(f1)
dic['auc']['ADASYN'].append(auc)
dic['gmean']['ADASYN'].append(gmean)
print('ADASYN building id transforming took %fs!' % (time.time() - start_time))
# ------------------------------------------------Safe-Level-SMOTE----------------------------------------------
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 训练
sb = SafeLevelSMOTE(N=100, k_neighbors=5, random_state=42)
# 预测
X_res, y_res = sb.fit_sample(X_train_minmax, y[train])
model = tree.DecisionTreeClassifier(max_depth=8, min_samples_split=10, random_state=42)
model.fit(X_res, y_res)
predict = model.predict(X_test_minmax)
probability = model.predict_proba(X_test_minmax)[:, 1]
precision = metrics.precision_score(y[test], predict)
recall = metrics.recall_score(y[test], predict)
if precision == 0:
f1 = 0
else:
f1 = 2 * (precision * recall) / (precision + recall)
auc = metrics.roc_auc_score(y[test], probability)
gmean = geometric_mean_score(y[test], predict)
dic['precision']['Safe-level'].append(precision)
dic['recall']['Safe-level'].append(recall)
dic['f1']['Safe-level'].append(f1)
dic['auc']['Safe-level'].append(auc)
dic['gmean']['Safe-level'].append(gmean)
print('Safe-level building id transforming took %fs!' % (time.time() - start_time))
# display
results.add_row(['CART',
np.mean(np.array(dic['precision']['CART'])),
np.mean(np.array(dic['recall']['CART'])),
np.mean(np.array(dic['auc']['CART'])),
np.mean(np.array(dic['f1']['CART'])),
np.mean(np.array(dic['gmean']['CART']))])
results.add_row(['SMOTE',
np.mean(np.array(dic['precision']['SMOTE'])),
np.mean(np.array(dic['recall']['SMOTE'])),
np.mean(np.array(dic['auc']['SMOTE'])),
np.mean(np.array(dic['f1']['SMOTE'])),
np.mean(np.array(dic['gmean']['SMOTE']))])
results.add_row(['Border1',
np.mean(np.array(dic['precision']['Border1'])),
np.mean(np.array(dic['recall']['Border1'])),
np.mean(np.array(dic['auc']['Border1'])),
np.mean(np.array(dic['f1']['Border1'])),
np.mean(np.array(dic['gmean']['Border1']))])
results.add_row(['Border2',
np.mean(np.array(dic['precision']['Border2'])),
np.mean(np.array(dic['recall']['Border2'])),
np.mean(np.array(dic['auc']['Border2'])),
np.mean(np.array(dic['f1']['Border2'])),
np.mean(np.array(dic['gmean']['Border2']))])
results.add_row(['ADASYN',
np.mean(np.array(dic['precision']['ADASYN'])),
np.mean(np.array(dic['recall']['ADASYN'])),
np.mean(np.array(dic['auc']['ADASYN'])),
np.mean(np.array(dic['f1']['ADASYN'])),
np.mean(np.array(dic['gmean']['ADASYN']))])
results.add_row(['Safe-level',
np.mean(np.array(dic['precision']['Safe-level'])),
np.mean(np.array(dic['recall']['Safe-level'])),
np.mean(np.array(dic['auc']['Safe-level'])),
np.mean(np.array(dic['f1']['Safe-level'])),
np.mean(np.array(dic['gmean']['Safe-level']))])
print(results)
def fetch(*args, **kwargs):
return fetch_datasets(*args, download_if_missing=True, **kwargs)
def test_fetch_error(filter_data, err_msg):
with pytest.raises(ValueError, match=err_msg):
fetch_datasets(filter_data=filter_data)
from imblearn.datasets import fetch_datasets
from sklearn.model_selection import train_test_split, cross_val_score, KFold
import numpy
datasets = [
'ecoli', 'optical_digits', 'satimage', 'pen_digits', 'abalone',
'sick_euthyroid', 'spectrometer', 'car_eval_34', 'isolet', 'us_crime',
'yeast_ml8', 'scene', 'libras_move', 'thyroid_sick', 'coil_2000',
'arrhythmia', 'solar_flare_m0', 'oil', 'car_eval_4', 'wine_quality',
'letter_img', 'yeast_me2', 'webpage', 'ozone_level', 'mammography',
'protein_homo', 'abalone_19'
]
for dataset in datasets:
object = fetch_datasets(data_home='./data/')[dataset]
X, y = object.data, object.target
train_X, test_X, train_y, test_y = train_test_split(
X, y) # splits 75%/25% by default
numpy.savez('./data/zendo_stable/' + dataset + '.npz',
train_X=train_X,
test_X=test_X,
train_y=train_y,
test_y=test_y)
horizontalalignment="center",
color="white" if cm[i, j] > thresh else "black",
)
ax.set_ylabel("True label")
ax.set_xlabel("Predicted label")
###############################################################################
# Load an imbalanced dataset
###############################################################################
# We will load the UCI SatImage dataset which has an imbalanced ratio of 9.3:1
# (number of majority sample for a minority sample). The data are then split
# into training and testing.
satimage = fetch_datasets()["satimage"]
X, y = satimage.data, satimage.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
stratify=y,
random_state=0)
###############################################################################
# Classification using a single decision tree
###############################################################################
# We train a decision tree classifier which will be used as a baseline for the
# rest of this example.
###############################################################################
# The results are reported in terms of balanced accuracy and geometric mean
# which are metrics widely used in the literature to validate model trained on
def run_eval(dataset, base_learners, methods):
if dataset == "wilt":
X, y, cl_names = load_wilt()
elif dataset == "adult":
X, y, cl_names = load_adult()
elif dataset == "diabetes":
X, y, cl_names = load_diabetes()
elif dataset == "phoneme":
X, y, cl_names = load_phoneme()
elif dataset == "mushroom":
X, y, cl_names = load_mushroom()
elif dataset == "electricity":
X, y, cl_names = load_electricity()
elif dataset == "speeddating":
X, y, cl_names = load_speed_dating()
elif dataset == "credit":
X, y, cl_names = load_credit()
elif dataset == "eeg_eye":
X, y, cl_names = load_eeg_eye()
elif dataset == "spam":
X, y, cl_names = load_spam()
elif dataset == "skin":
X, y, cl_names = load_skin()
elif dataset == "bank":
X, y, cl_names = load_bank()
elif dataset == "kdd":
X, y, cl_names = load_kdd()
elif dataset == "landsatM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "musk2":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "spliceM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "semeion_orig":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "waveformM":
X, y, cl_names = load_mat_data(dataset)
else:
from imblearn import datasets
data = datasets.fetch_datasets()[dataset]
cl_names = [
"feature_" + str(i) for i in range(0, data['data'].shape[1])
]
X = data['data']
y = data['target']
y[y != 1] = 0
processes = []
for method in methods:
p = Process(target=train_classifier,
args=(X, y, base_learners, method,
cl_names)) # Passing the list
p.start()
processes.append(p)
for p in processes:
p.join()
N = len(methods)
ind = numpy.arange(N) # the x locations for the groups
width = 0.35 # the width of the bars: can also be len(x) sequence
raw_data = dict()
for method in methods:
with open('temp_features/' + method, 'rb') as filehandle:
# read the data as binary data stream
model = pickle.load(filehandle)
# print (method, model.feature_importances_)
raw_data[method] = model.feature_importances_
f_num = len(model.feature_importances_)
index = ["Feature " + str(k) for k in range(1, f_num + 1)]
# index = ["Atrribute 1","Atrribute 2","Atrribute 3","Atrribute 4","Atrribute 5","Atrribute 6"]
df = pd.DataFrame(raw_data, index=index)
df = df.transpose()
ax = df.plot.bar(stacked=True, alpha=0.75, rot=25)
ax.set_ylabel("Feature importance")
ax.set_xlabel("Methods")
ax.legend(loc='center left', bbox_to_anchor=(0.1, 01.07),
ncol=3) # here is the magic
ax.figure.savefig('Images/features/' + dataset + '.png',
bbox_inches='tight',
dpi=200)
pass
if __name__ == '__main__':
import time
import prettytable
from collections import Counter
from sklearn import tree
from sklearn import metrics
from sklearn import preprocessing
from imblearn.datasets import fetch_datasets
from imblearn.metrics import geometric_mean_score
from sklearn.model_selection import StratifiedKFold, RepeatedStratifiedKFold
start_time = time.time()
dataset = fetch_datasets()['oil']
X = dataset.data
y = dataset.target
# print(Counter(y))
cv = StratifiedKFold(n_splits=5, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
dic = {'recall': [], 'precision': [], 'f1': [], 'auc': [], 'gmean': []}
results = prettytable.PrettyTable(["Classifier", "Precision", 'Recall', 'F-measure', 'AUC', 'G-mean'])
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 训练
sb = BorderSMOTE(N=100, m_neighbors=9, k_neighbors=5, random_state=42, kind='borderline1')
plt.xticks(tick_marks, classes, rotation=45)
plt.yticks(tick_marks, classes)
fmt = '.2f' if normalize else 'd'
thresh = cm.max() / 2.
for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
plt.text(j, i, format(cm[i, j], fmt),
horizontalalignment="center",
color="white" if cm[i, j] > thresh else "black")
plt.tight_layout()
plt.ylabel('True label')
plt.xlabel('Predicted label')
ozone = fetch_datasets()['ozone_level']
X, y = ozone.data, ozone.target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)
bagging = BaggingClassifier(random_state=0)
balanced_bagging = BalancedBaggingClassifier(random_state=0)
print('Class distribution of the training set: {}'.format(Counter(y_train)))
bagging.fit(X_train, y_train)
balanced_bagging.fit(X_train, y_train)
print('Class distribution of the test set: {}'.format(Counter(y_test)))
print('Classification results using a bagging classifier on imbalanced data')
y_pred_bagging = bagging.predict(X_test)
def obtain_data(dataset_name):
dataset = fetch_datasets()[dataset_name]
return dataset.data, dataset.target
def run_eval(dataset, folds, iterations, baseL, methods):
if dataset == "wilt":
X, y, cl_names = load_wilt()
elif dataset == "adult":
X, y, cl_names = load_adult()
elif dataset == "diabetes":
X, y, cl_names = load_diabetes()
elif dataset == "phoneme":
X, y, cl_names = load_phoneme()
elif dataset == "mushroom":
X, y, cl_names = load_mushroom()
elif dataset == "electricity":
X, y, cl_names = load_electricity()
elif dataset == "speeddating":
X, y, cl_names = load_speed_dating()
elif dataset == "credit":
X, y, cl_names = load_credit()
elif dataset == "eeg_eye":
X, y, cl_names = load_eeg_eye()
elif dataset == "spam":
X, y, cl_names = load_spam()
elif dataset == "skin":
X, y, cl_names = load_skin()
elif dataset == "bank":
X, y, cl_names = load_bank()
elif dataset == "kdd":
X, y, cl_names = load_kdd()
elif dataset == "landsatM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "musk2":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "spliceM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "semeion_orig":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "rain_aus":
X, y, cl_names = load_rain_aus()
elif dataset == "waveformM":
X, y, cl_names = load_mat_data(dataset)
else:
from imblearn import datasets
data = datasets.fetch_datasets()[dataset]
cl_names = ["feature_" + str(i) for i in range(0, data['data'].shape[1])]
X = data['data']
y = data['target']
y[y != 1] = 0
unique_attr = set([i.split("?")[0] for i in cl_names])
print(dataset + "\t" + str(len(unique_attr)) + "\t" + str(f'{sum(abs(y[y == 1])):,}') + "\t" + str(
f'{len(abs(y[y != 1])):,}') + "\t1:" + str(format(len(abs(y[y != 1])) / sum(y[y == 1]), '.2f')))
list_of_dicts = []
list_of_dicts_stats = []
for t_dict in range(0, len(methods)):
list_of_dicts.append(defaultdict(dict))
list_of_dicts_stats.append(defaultdict(dict))
for weak_learners in baseL:
for item in list_of_dicts:
item[weak_learners] = defaultdict(list)
for weak_learners in baseL:
for item in list_of_dicts_stats:
item[weak_learners] = defaultdict(list)
for samples in range(0, iterations):
sss = StratifiedKFold(n_splits=folds, shuffle=True, random_state=int(time.time()))
for weak_learners in baseL:
print("iteration=", samples, " weak learners=", weak_learners)
# for weak_learners in baseL:
for train_index, test_index in sss.split(X, y):
X_train, X_test = X[train_index], X[test_index]
y_train, y_test = y[train_index], y[test_index]
processes = []
for method in methods:
p = Process(target=train_and_predict,
args=(X_train, y_train, X_test, weak_learners, method, cl_names))
p.start()
processes.append(p)
for p in processes:
p.join()
for index, method in enumerate(methods):
with open('temp_preds/' + method, 'rb') as filehandle:
list_of_dicts[index] = update_performance_stats(
calculate_performance(y_test, pickle.load(filehandle)),
list_of_dicts[index],
weak_learners
)
with open('temp_preds/stats_' + method, 'rb') as filehandle:
list_of_dicts_stats[index] = update_resource_stats(pickle.load(filehandle),
list_of_dicts_stats[index],
weak_learners,
method
)
plot_single_dataset(methods, list_of_dicts, "Images/Performance/" + dataset + "/", baseL)
plot_resource_stats_time(methods, list_of_dicts_stats, "Images/Performance/" + dataset + "/Resource/", baseL)
plot_resource_stats_scores(methods, list_of_dicts_stats, "Images/Performance/" + dataset + "/Resource/", baseL)
return list_of_dicts, list_of_dicts_stats
plt.xticks(tick_marks, classes, rotation=45)
plt.yticks(tick_marks, classes)
fmt = '.2f' if normalize else 'd'
thresh = cm.max() / 2.
for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
plt.text(j, i, format(cm[i, j], fmt),
horizontalalignment="center",
color="white" if cm[i, j] > thresh else "black")
plt.tight_layout()
plt.ylabel('True label')
plt.xlabel('Predicted label')
ozone = fetch_datasets()['ozone_level']
X, y = ozone.data, ozone.target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)
bagging = BaggingClassifier(random_state=0)
balanced_bagging = BalancedBaggingClassifier(random_state=0)
print('Class distribution of the training set: {}'.format(Counter(y_train)))
bagging.fit(X_train, y_train)
balanced_bagging.fit(X_train, y_train)
print('Class distribution of the test set: {}'.format(Counter(y_test)))
print('Classification results using a bagging classifier on imbalanced data')
y_pred_bagging = bagging.predict(X_test)
return X_resampled, y_resampled
if __name__ == '__main__':
import time
import prettytable
from collections import Counter
from sklearn import tree
from sklearn import metrics
from sklearn import preprocessing
from imblearn.datasets import fetch_datasets
from imblearn.metrics import geometric_mean_score
from sklearn.model_selection import StratifiedKFold, RepeatedStratifiedKFold
start_time = time.time()
dataset = fetch_datasets()['satimage']
X = dataset.data
y = dataset.target
print(Counter(y))
cv = StratifiedKFold(n_splits=10, random_state=42, shuffle=True)
# cv = RepeatedStratifiedKFold(n_repeats=5, n_splits=10, random_state=42)
dic = {'recall': [], 'precision': [], 'f1': [], 'auc': [], 'gmean': []}
results = prettytable.PrettyTable(
["Classifier", "Precision", 'Recall', 'F-measure', 'AUC', 'G-mean'])
for train, test in cv.split(X, y):
# 预处理
scaler = preprocessing.MinMaxScaler().fit(X[train])
X_train_minmax = scaler.transform(X[train])
X_test_minmax = scaler.transform(X[test])
# 训练
def test_fetch_error(filter_data, err_msg):
with pytest.raises(ValueError, match=err_msg):
fetch_datasets(filter_data=filter_data)
def run_eval(dataset, baseL, methods):
if dataset == "wilt":
X, y, cl_names = load_wilt()
elif dataset == "adult":
X, y, cl_names = load_adult()
elif dataset == "diabetes":
X, y, cl_names = load_diabetes()
elif dataset == "phoneme":
X, y, cl_names = load_phoneme()
elif dataset == "mushroom":
X, y, cl_names = load_mushroom()
elif dataset == "electricity":
X, y, cl_names = load_electricity()
elif dataset == "speeddating":
X, y, cl_names = load_speed_dating()
elif dataset == "credit":
X, y, cl_names = load_credit()
elif dataset == "eeg_eye":
X, y, cl_names = load_eeg_eye()
elif dataset == "spam":
X, y, cl_names = load_spam()
elif dataset == "skin":
X, y, cl_names = load_skin()
elif dataset == "bank":
X, y, cl_names = load_bank()
elif dataset == "kdd":
X, y, cl_names = load_kdd()
elif dataset == "landsatM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "musk2":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "spliceM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "semeion_orig":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "waveformM":
X, y, cl_names = load_mat_data(dataset)
else:
from imblearn import datasets
data = datasets.fetch_datasets()[dataset]
cl_names = ["feature_" + str(i) for i in range(0, data['data'].shape[1])]
X = data['data']
y = data['target']
y[y != 1] = 0
print("===============-- " + dataset + " --===============")
processes = []
for method in methods:
p = Process(target=train_and_predict, args=(X, y, baseL, method))
p.start()
processes.append(p)
for p in processes:
p.join()
list_of_dicts = []
for method in methods:
with open('temp_preds_adaac/' + method, 'rb') as filehandle:
list_of_dicts.append(update_stats(pickle.load(filehandle)))
plot_amort_vs_non_amort(methods, list_of_dicts, baseL, "Images/Amort_vs_non_amort/" + dataset + "/")
return list_of_dicts
}
for data in dataset:
print("dataset : ", data)
'''
fetch_data = fetch_datasets()[data]
X = fetch_data.data
y = fetch_data.target
normalization_object = Normalizer()
X = normalization_object.fit_transform(X)
labelencoder = LabelEncoder()
y = labelencoder.fit_transform(y)
'''
fetch_data = fetch_datasets()[data]
X = fetch_data.data
y = fetch_data.target
Standard_object = StandardScaler()
X = Standard_object.fit_transform(X)
labelencoder = LabelEncoder()
y = labelencoder.fit_transform(y)
value, counts = np.unique(y, return_counts=True)
if counts[0]>= counts[1]:
fraction = int((counts[1]/counts[0])*100)
else:
from sklearn.model_selection import KFold
from imblearn.datasets import fetch_datasets
from photonai.base import Hyperpipe, PipelineElement, OutputSettings
from photonai.optimization import FloatRange, Categorical, IntegerRange
# example of imbalanced dataset
dataset = fetch_datasets()["coil_2000"]
X, y = dataset.data, dataset.target
# ratio class 0: 0.06%, class 1: 0.94%
my_pipe = Hyperpipe(
"basic_svm_pipe_no_performance",
optimizer="random_grid_search",
optimizer_params={"n_configurations": 10},
metrics=["accuracy", "precision", "recall"],
best_config_metric="recall",
outer_cv=KFold(n_splits=3),
inner_cv=KFold(n_splits=5),
verbosity=1,
output_settings=OutputSettings(project_folder="./tmp/"),
)
# ADD ELEMENTS TO YOUR PIPELINE
my_pipe += PipelineElement("StandardScaler")
my_pipe += PipelineElement(
"PCA", hyperparameters={"n_components": IntegerRange(5, 20)}, test_disabled=True
)
def run_eval(dataset, base_learners, methods):
if dataset == "wilt":
X, y, cl_names = load_wilt()
elif dataset == "adult":
X, y, cl_names = load_adult()
elif dataset == "diabetes":
X, y, cl_names = load_diabetes()
elif dataset == "phoneme":
X, y, cl_names = load_phoneme()
elif dataset == "mushroom":
X, y, cl_names = load_mushroom()
elif dataset == "electricity":
X, y, cl_names = load_electricity()
elif dataset == "speeddating":
X, y, cl_names = load_speed_dating()
elif dataset == "credit":
X, y, cl_names = load_credit()
elif dataset == "eeg_eye":
X, y, cl_names = load_eeg_eye()
elif dataset == "spam":
X, y, cl_names = load_spam()
elif dataset == "skin":
X, y, cl_names = load_skin()
elif dataset == "bank":
X, y, cl_names = load_bank()
elif dataset == "kdd":
X, y, cl_names = load_kdd()
elif dataset == "landsatM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "musk2":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "spliceM":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "semeion_orig":
X, y, cl_names = load_mat_data(dataset)
elif dataset == "waveformM":
X, y, cl_names = load_mat_data(dataset)
else:
from imblearn import datasets
data = datasets.fetch_datasets()[dataset]
cl_names = ["feature_" + str(i) for i in range(0, data['data'].shape[1])]
X = data['data']
y = data['target']
y[y != 1] = 0
list_of_scores = []
processes = []
for method in methods:
p = Process(target=train_classifier, args=(X, y, base_learners, method, cl_names)) # Passing the list
p.start()
processes.append(p)
for p in processes:
p.join()
for method in methods:
with open('temp/' + method, 'rb') as filehandle:
# read the data as binary data stream
list_of_scores.append(pickle.load(filehandle))
y[y != 1] = -1
for idx in range(0, len(list_of_scores)):
list_of_scores[idx] = numpy.array(list_of_scores[idx]) * y
overall_confs = []
positive_confs = []
negative_confs = []
for conf in list_of_scores:
overall_confs.append(conf)
positive_confs.append(conf[y == 1])
negative_confs.append(conf[y == -1])
num_bins = 40
fig, (ax1, ax2, ax3) = plt.subplots(nrows=1, ncols=3, figsize=(20, 4))
plt.rcParams.update({'font.size': 12})
colors = ['b', 'g', 'r', 'c', 'm', 'y', 'k', 'dimgray', 'peru', 'hotpink', 'tomato']
default_cycler = (cycler(color=colors) +
cycler(linestyle=['-', (0, (1, 1)), '--', '-.',
(0, (5, 10)),
(0, (5, 1)),
'-', (0, (1, 1)), '--', '-.',
(0, (5, 10))]))
ax1.set_prop_cycle(default_cycler)
ax2.set_prop_cycle(default_cycler)
ax3.set_prop_cycle(default_cycler)
ax1.set_title("Positive CDF")
ax1.grid(True)
ax1.set_xlim(-1, 1)
ax2.set_xlim(-1, 1)
ax3.set_xlim(-1, 1)
colors = ['b', 'g', 'r', 'c', 'm', 'y', 'k', 'dimgray', 'peru', 'hotpink', 'tomato', 'indigo', 'lightskyblue']
output = defaultdict(list)
for idx in range(0, len(positive_confs)):
pos_conf = positive_confs[idx]
counts_positives, bin_edges_positives = numpy.histogram(pos_conf, bins=num_bins, normed=True)
cdf_positives = numpy.cumsum(counts_positives)
# ax1.plot(bin_edges_positives[:-1], cdf_positives / cdf_positives[-1], label=methods[idx],color=colors[idx])
ax1.plot(bin_edges_positives[:-1], cdf_positives / cdf_positives[-1], label=methods[idx])
output[methods[idx]].append(bin_edges_positives[:-1])
output[methods[idx]].append(cdf_positives)
# ax1.legend(loc='best')
ax1.set_xlabel("Margin")
ax1.set_ylabel("Cumulative Distribution")
ax1.axhline(0, color='black')
ax1.axvline(0, color='black')
ax2.grid(True)
ax2.axhline(0, color='black')
ax2.axvline(0, color='black')
ax2.set_title("Negative CDF")
for idx in range(0, len(negative_confs)):
if idx == 0:
ax2.set_ylabel("Cumulative Distribution")
ax2.set_xlabel("Margin")
neg_conf = negative_confs[idx]
counts_negatives, bin_edges_negatives = numpy.histogram(neg_conf, bins=num_bins, normed=True)
cdf_negatives = numpy.cumsum(counts_negatives)
# ax2.plot(bin_edges_negatives[:-1], cdf_negatives / cdf_negatives[-1], label=methods[idx],color=colors[idx])
ax2.plot(bin_edges_negatives[:-1], cdf_negatives / cdf_negatives[-1], label=methods[idx])
output[methods[idx]].append(bin_edges_negatives[:-1])
output[methods[idx]].append(cdf_negatives)
ax3.grid(True)
ax3.axhline(0, color='black')
ax3.axvline(0, color='black')
ax3.set_title("Overall CDF")
for idx in range(0, len(negative_confs)):
if idx == 0:
ax3.set_ylabel("Cumulative Distribution")
ax3.set_xlabel("Margin")
over_conf = overall_confs[idx]
counts_overall, bin_edges_overall = numpy.histogram(over_conf, bins=num_bins, normed=True)
cdf_overall = numpy.cumsum(counts_overall)
# ax3.plot(bin_edges_overall[:-1], cdf_overall / cdf_overall[-1], label=methods[idx], color=colors[idx])
ax3.plot(bin_edges_overall[:-1], cdf_overall / cdf_overall[-1], label=methods[idx])
output[methods[idx]].append(bin_edges_overall[:-1])
output[methods[idx]].append(cdf_overall)
plt.legend(loc='upper center', bbox_to_anchor=(-0.7, 1.305), ncol=5)
if not os.path.exists("Images/cdf_plots/" + dataset):
os.makedirs("Images/cdf_plots/" + dataset)
plt.savefig("Images/cdf_plots/" + dataset + "/cdf_" + str(base_learners) + ".png", bbox_inches='tight',
dpi=200)
return output
from sklearn.model_selection import train_test_split, GridSearchCV from sklearn.pipeline import Pipeline from sklearn.preprocessing import StandardScaler from sklearn.neighbors import KNeighborsClassifier random_seed = 3 # ## Preparing the data # In[2]: np.random.seed(random_seed) # In[3]: libras = imb_datasets.fetch_datasets()['libras_move'] X, y = libras['data'], libras['target'] # In[4]: X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33) # ## Fitting a pipeline # In[5]: oversampler = sv.MulticlassOversampling(sv.distance_SMOTE()) classifier = KNeighborsClassifier(n_neighbors=5) # In[6]: