def experiment_invoke(dataset='bank'):
if dataset == 'wine':
reader = WineDataReader()
else:
reader = BankDataReader()
ds = DataSplitter(reader)
ds.read_split_data()
run_final = False
er = KMeansExp(reader, ds, run_final)
er.experiment()
er = EMExp(reader, ds, run_final)
er.experiment()
er = ANNExp(reader, ds)
er.experiment_bank()
er.experiment_clusters()
run_final = True
er = KMeansExp(reader, ds, run_final)
er.experiment()
er = EMExp(reader, ds, run_final)
er.experiment()
er = ANNExp(reader, ds)
er.experiment_bank()
er.experiment_clusters()
def __init__(self, convertor, splitter_method_index, splitter_method_parameter):
DataSplitter.__init__(self, convertor, splitter_method_index, splitter_method_parameter)
self.splitter_given_n = splitter_method_parameter
self.start_time_id = 0
self.methods = {
0: self.get_given_n_by_user,
1: self.get_given_n_by_item,
2: self.get_given_n_by_user_date,
3: self.get_given_n_by_item_date,
4: self.get_given_n_by_date,
}
def __init__(self, convertor, splitter_method_index,
splitter_method_parameter):
DataSplitter.__init__(self, convertor, splitter_method_index,
splitter_method_parameter)
self.splitter_given_n = splitter_method_parameter
self.start_time_id = 0
self.methods = {
0: self.get_given_n_by_user,
1: self.get_given_n_by_item,
2: self.get_given_n_by_user_date,
3: self.get_given_n_by_item_date,
4: self.get_given_n_by_date,
}
def splitClassificationTest(self):
train, test = DataSplitter().splitDataEqually(self.data, self.labelCol)
i = 0
for clf in self.classifiers:
self.testClassifier(clf, train, test, i)
i += 1
def splitMultiClassificationTest(self):
train, test = DataSplitter().splitDataEqually(self.data, self.labelCol)
i = 0
print("DATA IN TESTER", self.data)
for clf in self.classifiers:
self.testMultiClassifier(clf, train, test, i)
i += 1
def createFolders(self):
#it prepares the folders for k-fold cross validation
folders = [[]] * self.k
dataTmp = self.data.copy()
for u in range(0, self.k):
#it splits data equally according to each label
folder, dataTmp = DataSplitter().splitDataEqually(
dataTmp, self.labelCol, 1.0 / (self.k - u))
folders[u] = folder
self.folders = folders
data = data.iloc[:, 1:]
# One-hot encode the data using pandas get_dummies
data = pd.get_dummies(data)
# data transformation: real values into labels to classify
data = data.apply(transRow, axis=1)
notWeek = eliminateWeekSections(data.columns)
print(notWeek)
data = data[data.columns[notWeek]]
print(data.columns)
labelName = "shares"
train, test = DataSplitter().splitDataEqually(data, labelName)
Y_train = pd.factorize(train[labelName])[0]
X_train_origin = train.iloc[:, 0:train.columns.size - 1].copy()
Y_test = pd.factorize(test[labelName])[0]
X_test_origin = test.iloc[:, 0:test.columns.size - 1].copy()
scaler = preprocessing.MinMaxScaler(feature_range=(-1, 1))
scaler.fit(X_train_origin)
#scaling of training data
X_train_origin = pd.DataFrame(scaler.transform(X_train_origin.copy()), columns=X_train_origin.columns)
# apply same transformation to test data
X_test_origin = pd.DataFrame(scaler.transform(X_test_origin.copy()), columns=X_test_origin.columns)
trainTmp = X_train_origin.copy()
trainTmp[labelName] = Y_train
def __init__(self, convertor, splitter_method_index,
splitter_method_parameter):
DataSplitter.__init__(self, convertor, splitter_method_index,
splitter_method_parameter)
if(__name__ == "__main__"):
np.random.seed(12345)
# Read in data and display first 5 rows
data = pd.read_csv('regression.csv', sep=",")
print('The shape of our data is:', data.shape)
#one hot encoding: transorming nominal values
labelName = "G3"
labels = data[labelName]
data = pd.get_dummies(data)
data[labelName] = labels
train, test = DataSplitter().splitData(data.copy())
print(train.copy())
#preparing test and training for final evaluation: using copies not to create problems
scaler = preprocessing.MinMaxScaler(feature_range=(-1, 1))
#don't cheat: fit only on training data
scaler.fit(train)
trainTmp = pd.DataFrame(scaler.transform(train.copy()), columns=train.columns)
# apply same transformation to test data
testTmp = pd.DataFrame(scaler.transform(test.copy()), columns=test.columns)
fsSize = train.columns.size
data = data.iloc[:, 1:]
# One-hot encode the data using pandas get_dummies
data = pd.get_dummies(data)
# data transformation: real values into labels to classify
data = data.apply(transRow, axis=1)
notWeek = eliminateWeekSections(data.columns)
print(notWeek)
data = data[data.columns[notWeek]]
print(data.columns)
labelName = "shares"
train, test = DataSplitter().splitDataEqually(data, labelName)
Y_train = pd.factorize(train[labelName])[0]
X_train_origin = train.iloc[:, 0:train.columns.size - 1].copy()
Y_test = pd.factorize(test[labelName])[0]
X_test_origin = test.iloc[:, 0:test.columns.size - 1].copy()
scaler = preprocessing.MinMaxScaler(feature_range=(-1, 1))
scaler.fit(X_train_origin)
#scaling of training data
X_train_origin = pd.DataFrame(scaler.transform(X_train_origin.copy()),
columns=X_train_origin.columns)
# apply same transformation to test data
X_test_origin = pd.DataFrame(scaler.transform(X_test_origin.copy()),
columns=X_test_origin.columns)
def __init__(self, convertor, splitter_method_index, splitter_method_parameter):
DataSplitter.__init__(self, convertor, splitter_method_index, splitter_method_parameter)
data = data.iloc[:, 1:]
# print(data)
# One-hot encode the data using pandas get_dummies
data = pd.get_dummies(data)
notWeek = eliminateWeekSections(data.columns)
print(notWeek)
data = data[data.columns[notWeek]]
print(data.columns)
labelName = "shares"
train, test = DataSplitter().splitData(data.copy())
print("Splitted")
print("Fitting")
scaler = preprocessing.MinMaxScaler(feature_range=(-1, 1))
scaler.fit(train)
print("Fitted")
trainTmp = pd.DataFrame(scaler.transform(train.copy()),
columns=train.columns)
# apply same transformation to test data
testTmp = pd.DataFrame(scaler.transform(test.copy()), columns=test.columns)
def split_data(self, save_path, experiment_id):
if not self.load_train_test_data(save_path, experiment_id):
self.methods[self.splitter_method_index](self.splitter_given_n)
DataSplitter.split_data(self, save_path, experiment_id)
def split_data(self, save_path, experiment_id):
if not self.load_train_test_data(save_path, experiment_id):
self.methods[self.splitter_method_index](self.splitter_given_n)
DataSplitter.split_data(self, save_path, experiment_id)