Linear Regression Algorithm



Linear regression is more than 200 years old algorithm is used for predicting properties with a training data set.

In this blog we will learn

What is linear regression
Calculate statistical quantities from a training data set.
Calculate linear regression coefficients from a data set.
Make predictions using linear regression.
Use sklearn library to make predictions with linear regression

Linear Regression

Simple linear regression is a straight line equation between independent and dependent variables. That straight equation is

Here, y is a dependent variable on x (an independent variable). We will need to estimate slope and the y intercept from the training data set and once we get these coefficients we can use this equation any value for y given x as input. But why this straight line equation?

Suppose that we have this data for the per capita income of the US(in dollars) for the years 1970 to 2016.

I will represent the data using a jupyter notebook, and various python libraries such as pandas, numpy sklearn, matplotlib with an alias name.

import pandas as pd

import numpy as np

import matplotlib.pyplot as plt

and plot the available data(training data set) using a scatter plot diagram.

The first five rows and the 2 columns of the data is as follows

df = pd.read_csv(‘percapita.csv’)


df.head()   # first five rows of the file

 

 

year

per_capita

0

1970

3399.299037

1

1971

3768.297935

2

1972

4251.175484

3

1973

4804.463248

4

1974

5576.514583

Now plotting the above data with 46 columns and 2 rows

%inline matplotlib


plt.scatter(df.year,df.per_capita)

<matplotlib.collections.PathCollection at 0x7f3a4437e208>

Now there could be more than one line of equations which satisfies the condition for finding the regression or prediction values as.

But to find the best line which fits the regression with the least error value we will need to calculate the coefficients of the equation.

So to calculate these coefficients, you’ll need to calculate the mean of both the properties first, and then find their difference from mean.

plt.xlabel(‘years (1960 -2016)’)

plt.ylabel(‘per_capita in dollars’)

plt.scatter(df.year,df.per_capita)

plt.scatter(np.mean(df.year),np.mean(df.per_capita),color=’red’)

<matplotlib.collections.PathCollection at 0x7f3d8322e898>

Now to draw a relation between these points we will need a straight line equation using Least Square Method (to have the least difference between predicted line and the observed values).

So these coefficients can be calculated with

Here, (x – xฬ… )is the difference between the actual points of x and the mean value(1993.0) and (y-ศณ)is the difference between the actual value of y from the mean point (18920.1370).

year

per_capita_income(US$)

x-xฬ…

y-ศณ

(x-xฬ…)2

(y-ศณ)(x-xฬ…)

1970

3399.299037

23

-15,520.837963

529

-3,56,979.273149

1971

3768.297935

22

-15,151.839065

484

-3,33,340.45943

When you have calculated slope(m),in this case {828.46507522}  the equation for the mean value of x and y will be

         18920.1370 = {828.46507522}*1993.0 + c

Which on further calculation will give,

   c = -1632210.7578554575

So now the equation, for any point of value will be

y = {828.46507522}*x + {-1632210.7578554575}

And there you are to predict any value of per capita for a given year.

Check out this Online Data Science Course by Fita, which includes Supervised,Unsupervised machine learning algorithms,Data Analysis Manipulation and visualisation,reinforcement testing, hypothesis testing and much more to make an industry required data scientist at an affordable price, which includes certification, support with career guidance assistance.

Or with a python function it can be implemented as

#covariance between x and y

def covar(x,x_mean,y,y_mean):

covariance = 0.0

for i in range(len(y)):

covariance += (x[i] – x_mean) * (y[i] – y_mean)

return covar

 

#variance for difference between actual and mean value

def variance(values):

return np.var(values)

 

# slope and intercept

def coefficients(row_1,row_2):

x_mean, y_mean = np.mean(row_1), np.mean(row_2)

slope = covar(row_1,x_mean,row_2,y_mean)/variance(row_1)

intercept = x_mean – (slope * y_mean)

return [slope, intercept]

 

def simple_linear_regression(df,test_values):

predictions = []

m, x = coefficients(df[[‘years’]],df[[‘per_capita’]])

for i in test_values:

y_values = x + m * i

predictions.append(y_values)

return predictions

Estimate regression equation using sklearn

And now here’s how you would do it with python sklearn library.Import linear_model from the library and create an instance of it.

from sklearn import linear_model

reg = linear_model.LinerRegression()

 

  # passing per capita as a dependent variable on per capita


reg.fit(df[[‘years’]],df.per_capita)

Now the model is ready for a best fit equation line, we can find out the slope and the y intercept with reg.coef_ and reg.intercept_

reg.coef_

reg.intercept_

Which outputs

828.46507522

-1632210.7578554575

Now let us visualise the data with matplotlib

plt.xlabel(‘years (1960 -2020)’)

plt.ylabel(‘per_capita (in dollars)’)


plt.scatter(np.mean(df.year),np.mean(df.per_capita),color=’red’)


plt.plot(df.year,reg.predict(df[[‘year’]]),color=’black’)

[<matplotlib.lines.Line2D at 0x7fa1daffbba8>]

and then use the predict method to predict any value of per capita for a given year.

reg.predict([[2020]])

Output

41288.69409442

Now let’s predict the per capita for recent years(testing data set) ,and store them in a csv file

df_2 = pd.read_csv(‘years.csv’)

df_2.head()

 

 

year

0

2016

1

2017

2

2018

3

2019

4

2020

Now store the predicted values in the new column of the years.csv file.

predicts = reg.predict(df_2)


df_2[‘predicted_per_capita’] = predicts  # creating new column

df_2.to_csv(‘predictions.csv’)    # creating new file


df_2.head()    # first five rows of the file

 

 

year

predicted_per_Capita

0

2016

37974.833794

1

2017

38803.298869

2

2018

39631.763944

3

2019

40460.229019

4

2020

41288.694094

You might notice the difference between the actual value of 2016 and the predicted value of 2016. This is known as mean squared error, and the correctness of the equation can be found with the R Square Method also known as coefficient of determination or coefficient of multiple determination. This R2can be calculated with the following formula.

     R2=(yp-ศณ)(xp-xฬ… )

If the more the R2is less than 1 the more the values are far the regression line.

This was all about linear regression algorithm with an example of predicting per capita income of US for several years with a trained data set.To get in-depth knowledge of Python along with its various applications and real-time projects, you can enroll in Python Training in Chennai or Python Training in Bangalore by FITA or enroll for a Data science course at Chennai or Data science course in Bangalore which includes Supervised, Unsupervised machine learning algorithms, Data Analysis Manipulation and visualisation, reinforcement testing, hypothesis testing and much more to make an industry required data scientist at an affordable price, which includes certification, support with career guidance assistance.






Quick Enquiry

Please wait while submission in progress...


Contact Us

Chennai

  93450 45466

Bangalore

 93450 45466

Coimbatore

 95978 88270

Online

93450 45466

Madurai

97900 94102

Pondicherry

93635 21112

For Hiring

 93840 47472
 hr@fita.in

Corporate Training

 90036 23340


Read More Read less

FITA Academy Branches

Chennai

Bangalore

Coimbatore

Other Locations

FITA Academy - Velachery
Plot No 7, 2nd floor,
Vadivelan Nagar,
Velachery Main Road,
Velachery, Chennai - 600042
Tamil Nadu

    :   93450 45466

FITA Academy - Anna Nagar
No 14, Block No, 338, 2nd Ave,
Anna Nagar,
Chennai 600 040, Tamil Nadu
Next to Santhosh Super Market

    :   93450 45466

FITA Academy - T Nagar
05, 5th Floor, Challa Mall,
T Nagar,
Chennai 600 017, Tamil Nadu
Opposite to Pondy Bazaar Globus

    :   93450 45466

FITA Academy - Tambaram
Nehru Nagar, Kadaperi,
GST Road, West Tambaram,
Chennai 600 045, Tamil Nadu
Opposite to Saravana Jewellers Near MEPZ

    :   93450 45466

FITA Academy - Thoraipakkam
5/350, Old Mahabalipuram Road,
Okkiyam Thoraipakkam,
Chennai 600 097, Tamil Nadu
Next to Cognizant Thoraipakkam Office and Opposite to Nilgris Supermarket

    :   93450 45466

FITA Academy - Porur
17, Trunk Rd,
Porur
Chennai 600116, Tamil Nadu
Above Maharashtra Bank

    :   93450 45466

FITA Academy Marathahalli
No 7, J J Complex,
ITPB Road, Aswath Nagar,
Marathahalli Post,
Bengaluru 560037

    :   93450 45466

FITA Academy - Saravanampatty
First Floor, Promenade Tower,
171/2A, Sathy Road, Saravanampatty,
Coimbatore - 641035
Tamil Nadu

    :   95978 88270

FITA Academy - Singanallur
348/1, Kamaraj Road,
Varadharajapuram, Singanallur,
Coimbatore - 641015
Tamil Nadu

    :   95978 88270

FITA Academy - Madurai
No.2A, Sivanandha salai,
Arapalayam Cross Road,
Ponnagaram Colony,
Madurai - 625016, Tamil Nadu

    :   97900 94102

FITA Academy - Pondicherry
410, Villianur Main Rd,
Sithananda Nagar, Nellitope,
Puducherry - 605005
Near IG Square

    :   93635 21112

FITA Academy - Tiruppur
61D, Poongodi Towers 2nd floor,
Periyar Colony Bus Stop,
Tirupur - 641 652

    :   9940122502

Read More Read less
  • Are You Located in Any of these Areas

    Adyar, Adambakkam, Anna Salai, Ambattur, Ashok Nagar, Aminjikarai, Anna Nagar, Besant Nagar, Chromepet, Choolaimedu, Guindy, Egmore, K.K. Nagar, Kodambakkam, Koyambedu, Ekkattuthangal, Kilpauk, Meenambakkam, Medavakkam, Nandanam, Nungambakkam, Madipakkam, Teynampet, Nanganallur, Navalur, Mylapore, Pallavaram, Purasaiwakkam, OMR, Porur, Pallikaranai, Poonamallee, Perambur, Saidapet, Siruseri, St.Thomas Mount, Perungudi, T.Nagar, Sholinganallur, Triplicane, Thoraipakkam, Tambaram, Vadapalani, Valasaravakkam, Villivakkam, Thiruvanmiyur, West Mambalam, Velachery and Virugambakkam.

    FITA Velachery or T Nagar or Thoraipakkam OMR or Anna Nagar or Tambaram or Porur branch is just few kilometre away from your location. If you need the best training in Chennai, driving a couple of extra kilometres is worth it!