In [2]:

    

import graphlab


    

In [3]:

    

sales = graphlab.SFrame('home_data.gl/')


    

    




[INFO] graphlab.cython.cy_server: GraphLab Create v2.1 started. Logging: /tmp/graphlab_server_1489629213.log






    




This non-commercial license of GraphLab Create for academic use is assigned to y_xwang@163.com and will expire on March 13, 2018.

In [4]:

    

graphlab.canvas.set_target('ipynb')

sales.show(view="Scatter Plot", x="sqft_living", y="price")


    

In [22]:

    

train_data, test_data = sales.random_split(.8, seed=0)


    

In [23]:

    

sqft_model = graphlab.linear_regression.create(train_data, target='price', features=['sqft_living'])


    

    




PROGRESS: Creating a validation set from 5 percent of training data. This may take a while.
          You can set ``validation_set=None`` to disable validation tracking.







    





Linear regression:






    





--------------------------------------------------------






    





Number of examples          : 16498






    





Number of features          : 1






    





Number of unpacked features : 1






    





Number of coefficients    : 2






    





Starting Newton Method






    





--------------------------------------------------------






    





+-----------+----------+--------------+--------------------+----------------------+---------------+-----------------+






    





| Iteration | Passes   | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |






    





+-----------+----------+--------------+--------------------+----------------------+---------------+-----------------+






    





| 1         | 2        | 0.043036     | 4336365.999378     | 2154015.621817       | 262672.861452 | 267995.181634   |






    





+-----------+----------+--------------+--------------------+----------------------+---------------+-----------------+






    





SUCCESS: Optimal solution found.






    

In [24]:

    

print sqft_model.evaluate(test_data)


    

    




{'max_error': 4133131.4124645432, 'rmse': 255214.74371849818}

In [25]:

    

import matplotlib.pyplot as plt
%matplotlib inline


    

In [26]:

    

plt.plot(test_data['sqft_living'], test_data['price'], '.', test_data['sqft_living'], sqft_model.predict(test_data), '-')


    

    
Out[26]:





[<matplotlib.lines.Line2D at 0x7fd5fefd1d10>,
 <matplotlib.lines.Line2D at 0x7fd5fefd1dd0>]






    

In [27]:

    

print sqft_model.get('coefficients')


    

    




+-------------+-------+---------------+---------------+
|     name    | index |     value     |     stderr    |
+-------------+-------+---------------+---------------+
| (intercept) |  None | -49223.975163 | 5048.50707157 |
| sqft_living |  None | 283.224728281 |  2.2205869952 |
+-------------+-------+---------------+---------------+
[2 rows x 4 columns]

In [33]:

    

my_features = ['bedrooms', 'bathrooms', 'sqft_living', 'sqft_lot', 'floors', 'zipcode']


    

In [34]:

    

sales[my_features].show()


    

In [36]:

    

my_features_model = graphlab.linear_regression.create(train_data, target='price', features=my_features)


    

    




PROGRESS: Creating a validation set from 5 percent of training data. This may take a while.
          You can set ``validation_set=None`` to disable validation tracking.







    





Linear regression:






    





--------------------------------------------------------






    





Number of examples          : 16478






    





Number of features          : 6






    





Number of unpacked features : 6






    





Number of coefficients    : 115






    





Starting Newton Method






    





--------------------------------------------------------






    





+-----------+----------+--------------+--------------------+----------------------+---------------+-----------------+






    





| Iteration | Passes   | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |






    





+-----------+----------+--------------+--------------------+----------------------+---------------+-----------------+






    





| 1         | 2        | 0.088427     | 3754563.110236     | 1440366.489067       | 181991.462839 | 182897.196164   |






    





+-----------+----------+--------------+--------------------+----------------------+---------------+-----------------+






    





SUCCESS: Optimal solution found.






    

In [37]:

    

print my_features_model.evaluate(test_data)


    

    




{'max_error': 3500622.0565247936, 'rmse': 179853.96603646813}