Tests an ambiguous problem with Linear Regression

test_ambiguous_lr.py

import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
import numpy as np

NSAMPLE=5000
x_data = np.float32(np.random.uniform(-10.5, 10.5, (1, NSAMPLE))).T
r_data = np.float32(np.random.normal(size=(NSAMPLE,1)))
y_data = np.float32(np.sin(0.75*x_data)*7.0+x_data*0.5+r_data*1.0)
x_data, y_data = y_data, x_data #swap x and y
X = x_data
y = (y_data>5).astype(np.float)

positive_indexes = y==1
negative_indexes = y==0
plt.plot(x_data[negative_indexes], y_data[negative_indexes], '.', label='Negative')
plt.plot(x_data[positive_indexes], y_data[positive_indexes], '.', label='Positive')
plt.legend(loc="best")
plt.title("Real")

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1)
regularization = 1e60
class_weight = {0:1, 1:2}
clf = LogisticRegression(solver='liblinear', tol=1e-10, max_iter=10000, C=regularization, class_weight=class_weight);
clf.fit(X_train, y_train.ravel());
y_predicted = clf.predict(X)
print "Coeff: {}, Intercept: {}".format(clf.coef_, clf.intercept_)
print "Score over training: ", clf.score(X_train, y_train)
print "Score over testing: ", clf.score(X_test, y_test)
print "Score total: ", clf.score(X, y)
print "Real percent of class 1: ", sum(y)/len(y)
print "Predicted percent of class 1: ", float(sum(y_predicted==1))/len(y_predicted)
positive_indexes = y_predicted==1
negative_indexes = y_predicted==0
plt.figure()
plt.plot(x_data[negative_indexes], y_data[negative_indexes], '.', label='Negative')
plt.plot(x_data[positive_indexes], y_data[positive_indexes], '.', label='Positive')
plt.legend(loc="best")
plt.title("Classifier")
plt.show()