def ex_dark_light_scatter_sns():
X, Y = make_regression(n_samples=100, n_features=2, noise=50.0)
Y[Y <= 0] = 0
Y[Y > 0] = 1
P = tools_plot_v2.Plotter(folder_out, dark_mode=False)
P.plot_2D_features_multi_Y(X, Y, filename_out='seaborn_scatter_light.png')
P = tools_plot_v2.Plotter(folder_out,dark_mode=True)
P.plot_2D_features_multi_Y(X, Y, filename_out='seaborn_scatter_dark.png')
return
def ex_dark_light_line():
tpr = numpy.linspace(0,1,20)
fpr = numpy.linspace(0.3,0.9,20)
P = tools_plot_v2.Plotter(folder_out, dark_mode=False)
P.plot_tp_fp(tpr, fpr, 0.5, caption='', filename_out='matplotlib_line_light.png')
P = tools_plot_v2.Plotter(folder_out, dark_mode=True)
P.plot_tp_fp(tpr, fpr, 0.5, caption='', filename_out='matplotlib_line_dark.png')
return
def __init__(self,Classifier,folder_out=None,dark_mode=False):
self.classifier = Classifier
self.P = tools_plot_v2.Plotter(folder_out,dark_mode)
self.folder_out = folder_out
if folder_out is not None and (not os.path.exists(folder_out)):
os.mkdir(folder_out)
return
def ex_train_test(X,Y):
C = classifier_LM.classifier_LM()
ML = tools_ML_v2.ML(C, folder_out + 'original/')
P = tools_plot_v2.Plotter(folder_out+'original/')
df = pd.DataFrame(data=(numpy.hstack((Y.reshape((-1, 1)), X))),columns=['target'] + ['%d' % c for c in range(X.shape[1])])
P.pairplots_df(df, idx_target=0)
ML.E2E_train_test_df(df,idx_target=0)
ML = tools_ML_v2.ML(C, folder_out + 'sampled/')
P = tools_plot_v2.Plotter(folder_out + 'sampled/')
X_Sampled, Y_Sampled = get_SMOTE_UnderSampler(X,Y,do_debug=True)
df_sampled = pd.DataFrame(data=(numpy.hstack((Y_Sampled.reshape((-1, 1)), X_Sampled))),columns=['target'] + ['%d' % c for c in range(X.shape[1])])
ML.E2E_train_test_df(df_sampled, idx_target=0)
P.pairplots_df(df_sampled, idx_target=0)
return
def __init__(self, app, folder_out, dark_mode):
self.folder_out = folder_out
self.P = tools_plot_v2.Plotter(folder_out, dark_mode)
self.app = app
self.filename_retro_df = 'retro.csv'
self.TS = tools_TS.tools_TS(
Classifier=TS_AutoRegression.TS_AutoRegression(folder_out),
dark_mode=dark_mode,
folder_out=folder_out)
self.clear_cache()
return
def ex_random():
X, Y = make_regression(n_samples=1250, n_features=3, noise=50.0)
Y[Y <= 0] = 0
Y[Y > 0] = 1
C = classifier_LM.classifier_LM()
P = tools_plot_v2.Plotter(folder_out)
df = pd.DataFrame(data=(numpy.hstack((Y.reshape((-1, 1)), X))),columns=['target'] + ['%d' % c for c in range(X.shape[1])])
ML = tools_ML_v2.ML(C, folder_out)
ML.E2E_train_test_df(df,idx_target=0)
P.pairplots_df(df, idx_target=0)
return
def ex_titanic():
#C = classifier_KNN.classifier_KNN()
#C = classifier_DTree.classifier_DT(folder_out=folder_out)
#C = classifier_RF.classifier_RF()
#C = classifier_Ada.classifier_Ada()
C = classifier_LM.classifier_LM()
P = tools_plot_v2.Plotter(folder_out)
df,idx_target = pd.read_csv(folder_in+'dataset_titanic.csv', sep='\t'),0
df.drop(labels=['alive', 'deck'], axis=1, inplace=True)
ML = tools_ML_v2.ML(C)
ML.E2E_train_test_df(df,idx_target=idx_target)
P.pairplots_df(df, idx_target=idx_target)
return
import numpy as numpy
import math
from sklearn import linear_model
from sklearn.datasets import make_regression
from sklearn.model_selection import train_test_split
from sklearn.metrics import r2_score
import pandas as pd
from sklearn import metrics
# ----------------------------------------------------------------------------------------------------------------------
import tools_plot_v2
# ----------------------------------------------------------------------------------------------------------------------
folder_in = './data/ex_datasets/'
folder_out = './data/output/'
# ----------------------------------------------------------------------------------------------------------------------
P = tools_plot_v2.Plotter(folder_out)
# ----------------------------------------------------------------------------------------------------------------------
def get_data_v1(filename_in, idx_target=0):
df = pd.read_csv(filename_in, sep='\t')
columns = df.columns.to_numpy()
idx = numpy.delete(numpy.arange(0, len(columns)), idx_target)
df_train, df_test = train_test_split(df.dropna(),
test_size=0.5,
shuffle=True)
X_train, Y_train = df_train.iloc[:, idx].to_numpy(
), df_train.iloc[:, [idx_target]].to_numpy()
X_test, Y_test = df_test.iloc[:, idx].to_numpy(
), df_test.iloc[:, [idx_target]].to_numpy()
import numpy
import pandas as pd
import matplotlib.pyplot as plt
from statsmodels.tsa.seasonal import seasonal_decompose
from pandas.plotting import autocorrelation_plot
from pandas.plotting import lag_plot
import statsmodels.api as sm
# ----------------------------------------------------------------------------------------------------------------------
import tools_plot_v2
# ----------------------------------------------------------------------------------------------------------------------
folder_in = './data/ex_TS/'
folder_out = './data/output/'
# ----------------------------------------------------------------------------------------------------------------------
P = tools_plot_v2.Plotter(folder_out, dark_mode=True)
# ----------------------------------------------------------------------------------------------------------------------
def ex_decompose(ts):
dates = numpy.array('2000-01-01', dtype=numpy.datetime64) + numpy.arange(
ts.shape[0])
df = pd.DataFrame({'data': ts.to_numpy()}, index=dates)
plt.clf()
plt.rcParams.update({'figure.figsize': (12, 7)})
seasonal_decompose(
df,
model='multiplicative').plot().suptitle('Multiplicative Decomposition')
plt.tight_layout()
plt.savefig(folder_out + 'decompose2_mult.png')
def __init__(self,Classifier=None,dark_mode=False,folder_out=None):
self.classifier = Classifier
self.Plotter = tools_plot_v2.Plotter(folder_out,dark_mode=dark_mode)
self.folder_out = folder_out
return
import cv2
import numpy as numpy
import pandas as pd
from sklearn.impute import SimpleImputer
# ----------------------------------------------------------------------------------------------------------------------
import tools_plot_v2
# ----------------------------------------------------------------------------------------------------------------------
folder_in = './data/ex_datasets/'
folder_out = './data/output/'
# ----------------------------------------------------------------------------------------------------------------------
P = tools_plot_v2.Plotter()
# ----------------------------------------------------------------------------------------------------------------------
def ex_is_missing(df):
A = (df.isnull()).to_numpy()
cv2.imwrite(folder_out + 'nans_1.png', 255 * A)
print(df)
print()
print(A)
return
# ----------------------------------------------------------------------------------------------------------------------
def ex_replace(df):
dct_replace = {numpy.NaN: 999.0}
print(df)
print()
df.replace(dct_replace, inplace=True)
print(df)
return