def linear_regression_example():
df = pd.read_csv('datasets/test.csv')
x = df.iloc[1:, 0].values
y = df.iloc[1:, 1].values
fig = plt.figure()
ax = fig.add_subplot(111)
ax.scatter(x[:], y[:], color='blue', marker='o', label='diamonds')
regession = LinearRegression()
regession.fit(x, y)
fit_x = np.linspace(np.min(x), np.max(x), 100)
fit_y = regession.get_slope() * fit_x + regession.get_intercept()
min_y = regession.get_slope() * fit_x + regession.get_intercept() - regession.get_interval()
max_y = regession.get_slope() * fit_x + regession.get_intercept() + regession.get_interval()
ax.plot(fit_x, fit_y, 'g--')
ax.plot(fit_x, min_y, 'r-')
ax.plot(fit_x, max_y, 'r-')
plt.show()
def linear_regression_example():
df = pd.read_csv(
'https://raw.githubusercontent.com/mwaskom/seaborn-data/master/diamonds.csv'
)
df = df[df['price'] > 1]
x = df.iloc[1:, 6].values
y = df.iloc[1:, 0].values
print(y[0])
fig = plt.figure()
ax = fig.add_subplot(111)
ax.scatter(x[:], y[:], color='blue', marker='o', label='diamonds')
ax.legend()
regession = LinearRegression()
regession.fit(x, y)
fit_x = np.linspace(np.min(x), np.max(x), 100)
fit_y = regession.get_slope() * fit_x + regession.get_intercept()
ax.plot(fit_x, fit_y)
plt.show()
from utils import *
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import axes3d
import pandas as pd
# Linear Regression example with randomly generated data
print("Running Linear Regression Demo")
print("\tLinear Regression Example on Randomly Generated Data")
X, Y = genPointsFittedToLine(np.array([-2]), 10)
plotTitle = "Linear Regression Demo:\n"
plotTitle += "Data Visualization for Randomly Generated Data Fitted to a Line"
plotRegression(X, Y, title=plotTitle)
# Create the Model
model = LinearRegression()
# Fit the model
model.fit(X, Y)
print("\t\tSquared Error: ", end="")
print(model.squaredError(X, Y))
# Display regression plane
plotTitle = "Regression Line Determined by Linear Regression Model for the\n"
plotTitle += "Randomly Generated Data"
plotRegression(X, Y, w=model.getWeights(), title=plotTitle)
# Linear Regression example with the Iris Plants Databse
print("\tLinear Regression Example on the Iris Plants Database")
# Download and read the data
df = pd.read_csv(
'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data',
def demo_helper(X, Y, learning_rate, iterations, title, x_label, y_label, x_lim=None, y_lim=None):
# Splitting the Dataset
X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size= 0.33, random_state= 101)
# Custom ML linear regression module
ml_regressor = LinearRegression(learning_rate=learning_rate, iterations=iterations)
# scikit-learn linear regression module
scikit_regressor = skLinearRegression()
# Perform linear regression using both scikit-learn and ML.py library
ml_regressor.fit(X_train, y_train) # custom version
scikit_regressor.fit(X_train, y_train) # scikit-learn version
# Make predictions
mlY_pred = ml_regressor.predict(X_test)
skY_pred = scikit_regressor.predict(X_test)
# Plot the X, Y data
plt.scatter(X, Y)
plt.title(title + ": Raw Data")
plt.xlabel(x_label)
plt.ylabel(y_label)
plt.show()
# Plot the X, Y_pred data
custom_label = title + ": ML.py Linear Regression Module\nLearning Rate: " + \
str(learning_rate) + ", Iterations: " + str(iterations)
plot_regression_line(mlY_pred, X_test, y_test, x_label, y_label,
custom_label, x_range=x_lim, y_range=y_lim)
plot_regression_line(skY_pred, X_test, y_test, x_label, y_label,
title + ": scikit-learn Linear Regression Module",
x_range=x_lim, y_range=y_lim)
# Evaluate Each Model's Performance
d_ml = {
'Metrics': ['Mean Absolute Error:',
'Mean Squared Error:',
'Mean Root Squared Error:'],
'Values': [mean_absolute_error(y_test, mlY_pred),
mean_squared_error(y_test, mlY_pred),
np.sqrt(mean_squared_error(y_test, mlY_pred))]
}
d_sk = {
'Metrics': ['Mean Absolute Error:',
'Mean Squared Error:',
'Mean Root Squared Error:'],
'Values': [mean_absolute_error(y_test, skY_pred),
mean_squared_error(y_test, skY_pred),
np.sqrt(mean_squared_error(y_test, skY_pred))]
}
df_ml = pd.DataFrame(data=d_ml)
df_sk = pd.DataFrame(data=d_sk)
fig, ax = plt.subplots(2)
cell_text = []
for row in range(len(df_ml)):
cell_text.append(df_ml.iloc[row])
ax[0].set_title(title + ": ML.py Linear Regression Module\nLearning Rate: " +
str(learning_rate) + ", Iterations: " + str(iterations))
ax[0].table(cellText=cell_text, colLabels=df_ml.columns, loc='center')
ax[0].axis(False)
cell_text = []
for row in range(len(df_sk)):
cell_text.append(df_sk.iloc[row])
ax[1].set_title(title + ": scikit-learn Linear Regression Module")
ax[1].table(cellText=cell_text, colLabels=df_sk.columns, loc='center')
ax[1].axis(False)
plt.show()
print("ML.py Linear Regression Weights for y = mx + b (" + title + "):")
print("Slope:", ml_regressor.slope)
print("Intercept:", ml_regressor.intercept, "\n")
print y
def all_func(x, y, model, state, window_state):
if draw and model:
linear_regression.plot(x, y, model.w)
if __name__ == "__main__":
i = 0
w = np.random.rand(num_features + 1, 1) * 2 - 1
w *= 5
m = LinearRegression.LinearRegressionStream(draw=False,
output=output,
alpha=0.001)
x = Stream('x')
linear_regression.init_plot()
model = Stream_Learn(data_train=x,
data_out=x,
train_func=m.train,
predict_func=m.predict,
min_window_size=min_window_size,
max_window_size=max_window_size,
step_size=step_size,
num_features=num_features,
all_func=all_func)
y = model.run()
def test_linearRegression():
df = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data', header=None)
print("Creating a weak positive correlation data set")
y = df.iloc[0:50, 0]
x = df.iloc[0:50, 1]
print("Creating Linear-Regression objet...")
linear = LinearRegression()
print("Predicting line of best fit given a weak positive correlation...")
linear.predict(x, y)
print("Graphing weak positive correlation...")
linear.graph(x, y)
print("Creating a random strong negative correlation dataset...")
x = np.random.randn(50)
y = np.random.randn(1) - 2 * x
print("Graphing weak positive correlation vs strong negative correlation dataset...")
linear.graph(x, y)
print("Retraining the linear regression leaner for a strong positive correlation dataset...")
linear.predict(x, y)
print("Graphing strong negative correlation...")
linear.graph(x, y)
print("Creating a weak negative correlation...")
y = df.iloc[0:50, 0]
x = df.iloc[0:50, 1]
np.negative(x)
print("Retraining the linear regression learner for a weak negative correlation dataset...")
linear.predict(x, y)
print("Graphing weak negative correlation...")
linear.graph(x, y)
print("Making a uniform dataset...")
x = np.random.rand(50)
print("Retraining the linear regression learner for a uniform distribution...")
linear.predict(x, y)
print("Graphing uniform distribution...")
linear.graph(x, y)