Course 2 week 1 lecture notebook 01¶

Create a Linear Model¶

Linear model using scikit-learn¶

We'll practice using a scikit-learn model for linear regression. You will do something similar in this week's assignment (but with a logistic regression model).

sklearn.linear_model.LinearRegression()

First, import LinearRegression, which is a Python 'class'.

In [ ]:

# Import the module 'LinearRegression' from sklearn
from sklearn.linear_model import LinearRegression

Next, use the class to create an object of type LinearRegression.

In [ ]:

# Create an object of type LinearRegression
model = LinearRegression()
model

Generate some data by importing a module 'load_data', which is implemented for you. The features in `X' are:

Age: (years)
Systolic_BP: Systolic blood pressure (mmHg)
Diastolic_BP: Diastolic blood pressure (mmHg)
Cholesterol: (mg/DL)

The labels in y indicate whether the patient has a disease (diabetic retinopathy).

y = 1 : patient has retinopathy.
y = 0 : patient does not have retinopathy.

In [ ]:

# Import the load_data function from the utils module
from utils import load_data

In [ ]:

# Generate features and labels using the imported function
X, y = load_data(100)

Explore the data by viewing the features and the labels

In [ ]:

# View the features
X.head()

In [ ]:

# Plot a histogram of the Age feature
X['Age'].hist();

In [ ]:

# Plot a histogram of the systolic blood pressure feature
X['Systolic_BP'].hist();

In [ ]:

# Plot a histogram of the diastolic blood pressure feature
X['Diastolic_BP'].hist();

In [ ]:

# Plot a histogram of the cholesterol feature
X['Cholesterol'].hist();

Also take a look at the labels

In [ ]:

# View a few values of the labels
y.head()

In [ ]:

# Plot a histogram of the labels
y.hist();

Fit the LinearRegression using the features in X and the labels in y. To "fit" the model is another way of saying that we are training the model on the data.

In [ ]:

# Fit the linear regression model
model.fit(X, y)
model

View the coefficients of the trained model.
The coefficients are the 'weights' or $\beta$ s associated with each feature
You'll use the coefficients for making predictions.

$\hat{y} = \beta_1x_1 + \beta_2x_2 + ... \beta_N x_N$

In [ ]:

# View the coefficients of the model
model.coef_

In the assignment, you will do something similar, but using a logistic regression, so that the output of the prediction will be bounded between 0 and 1.

This is the end of this practice section.¶

Please continue on with the lecture videos!