Upload linear_model.py
Browse files- linear_model.py +47 -0
linear_model.py
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Code source: Jaques Grobler
|
2 |
+
# License: BSD 3 clause
|
3 |
+
|
4 |
+
import matplotlib.pyplot as plt
|
5 |
+
import numpy as np
|
6 |
+
|
7 |
+
from sklearn import datasets, linear_model
|
8 |
+
from sklearn.metrics import mean_squared_error, r2_score
|
9 |
+
|
10 |
+
# Load the diabetes dataset
|
11 |
+
diabetes_X, diabetes_y = datasets.load_diabetes(return_X_y=True)
|
12 |
+
|
13 |
+
# Use only one feature
|
14 |
+
diabetes_X = diabetes_X[:, np.newaxis, 2]
|
15 |
+
|
16 |
+
# Split the data into training/testing sets
|
17 |
+
diabetes_X_train = diabetes_X[:-20]
|
18 |
+
diabetes_X_test = diabetes_X[-20:]
|
19 |
+
|
20 |
+
# Split the targets into training/testing sets
|
21 |
+
diabetes_y_train = diabetes_y[:-20]
|
22 |
+
diabetes_y_test = diabetes_y[-20:]
|
23 |
+
|
24 |
+
# Create linear regression object
|
25 |
+
regr = linear_model.LinearRegression()
|
26 |
+
|
27 |
+
# Train the model using the training sets
|
28 |
+
regr.fit(diabetes_X_train, diabetes_y_train)
|
29 |
+
|
30 |
+
# Make predictions using the testing set
|
31 |
+
diabetes_y_pred = regr.predict(diabetes_X_test)
|
32 |
+
|
33 |
+
# The coefficients
|
34 |
+
print("Coefficients: \n", regr.coef_)
|
35 |
+
# The mean squared error
|
36 |
+
print("Mean squared error: %.2f" % mean_squared_error(diabetes_y_test, diabetes_y_pred))
|
37 |
+
# The coefficient of determination: 1 is perfect prediction
|
38 |
+
print("Coefficient of determination: %.2f" % r2_score(diabetes_y_test, diabetes_y_pred))
|
39 |
+
|
40 |
+
# Plot outputs
|
41 |
+
plt.scatter(diabetes_X_test, diabetes_y_test, color="black")
|
42 |
+
plt.plot(diabetes_X_test, diabetes_y_pred, color="blue", linewidth=3)
|
43 |
+
|
44 |
+
plt.xticks(())
|
45 |
+
plt.yticks(())
|
46 |
+
|
47 |
+
plt.show()
|