Use sklearn for Linear Regression in Python

Introduction

We can use sklearn.linear_model.LinearRegression to do linear regression in Python. The following is the core syntax of using sklearn.

lm.fit(IVs, DV)

Where,

IVs: the independent variables

DV: the dependent variable

Example for Linear Regression Model

The following is the linear regression model, including household income as IVs and purchase intention as DV.

\[f(x)=b_0 +b_1 \times Price+b_2 \times Household \ Income \]

The following is the hypothetical data, including purchase intention as DV and prices and household income as IVs.

Purchase
Intention
PricesHousehold
Income
757
665
574
586
393
4103
Data for linear regression in Python

Step 1: Prepare the data

The following Python code generates the hypothetical data and then changed it into appropriate format.

import numpy as np
from sklearn.linear_model import LinearRegression
lm = LinearRegression()

# Input hypothetical data
Purchase_Intention=(7,6,5,5,3,4)

Prices=(5,6,7,8,9,10)
Household_income=(7,5,4,6,3,3)


# change it into the format we want
DV= np.transpose([Purchase_Intention])
print("DV: \n",DV)

# change it into the format we want
IVs=np.concatenate(([Prices], [Household_income]), axis=0)
IVs=np.transpose(IVs)
print("IVs: \n",IVs)

Output:

DV: 
 [[7]
 [6]
 [5]
 [5]
 [3]
 [4]]
IVs: 
 [[ 5  7]
 [ 6  5]
 [ 7  4]
 [ 8  6]
 [ 9  3]
 [10  3]]

Step 2: Use lm from sklearn.linear_model

.The following is the Python code of adding the IVs and DV in the lm.fit().

# apply sklearn.linear_model
result = lm.fit(IVs, DV)
print("Result is as follows:")
print("Intercept:\n",result.intercept_)
print("Regression Coefficients:\n", result.coef_)

Output:

Result is as follows:
Intercept:
 [6.73880597]
Regression Coefficients:
 [[-0.44776119  0.34701493]]

Step 3: Write out the regression model

We can write out a model with the estimated regression coefficients.

\[f(x)=6.73-0.45 \ Price+0.35 \ Household \ Income \]

We can see that price is negatively related with purchase intention, while household income is positively related with purchase intention.


Further Reading