Polynomial Regression
Dr. Singara Singh Kasana
Associate Professor
Computer Science and Engineering Department
Thapar Institute of Engineering and Technology
Patiala, Punjab
Polynomial Regression
Simple and multiple linear regressions are used when
the dependent(input) feature changes linearly with
respect to independent feature(s).
In some real time applications, dependent feature
may not change linearly with the changes in the
independent feature(s)
For such type of datasets, we can’t use simple or
multiple linear regressions. We need to find suitable
regression that fits with the dataset.
Representation of Different Regression Models
Simple Linear Regression
Y=𝛽0 + 𝛽1 ∗ 𝑥1
Multiple Linear Regression
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥2+. . . +𝛽𝑛 𝑥n
Polynomial Regression
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12+. . . +𝛽𝑛 𝑥 1n
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12+. . . +𝛽𝑛 𝑥 1n
If the degree of polynomial is 2, then n=2 and polynomial will be
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12
If the degree of polynomial is 3, then n=3 and polynomial will be
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12 + 𝛽3 ∗ 𝑥13
If the degree of polynomial is N, then n=N and polynomial will be
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12 + 𝛽3 ∗ 𝑥13 + . . . 𝛽𝑁 ∗ 𝑥1𝑁
Level Basic Salary(in Rs)
1 18000
2 19900
3 21700
4 25500
5 29200
6 35400
7 44900
8 47600
9 53100
10 56100
11 67700
12 78800
13 118500
14 131100
15 144200
16 182200
17 205400
18 225000
19 250000
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12 + 𝛽3 ∗ 𝑥13
Y=𝛽0 + 𝛽1 ∗ 𝑥1 + 𝛽2 ∗ 𝑥12 + 𝛽3 ∗ 𝑥13 + 𝛽4 ∗ 𝑥14
Implementation of Polynomial Regression
Step 1: Importing the required libraries
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
Implementation of Polynomial Regression
Step2: Load the dataset into dataframe
dataset = pd.read_csv('7thPayMatrix.csv')
print(dataset)
X=dataset.iloc[:,0]
Y=dataset.iloc[:,1:2]
print(X)
print(Y)
plt.scatter(X,Y)
plt.xlabel('Level')
plt.ylabel('Salary')
plt.show()
Polynomial Linear Regression
Step 3: Pre-process the independent feature using a particular degree polynomial
from sklearn.preprocessing import PolynomialFeatures
Obj = PolynomialFeatures(degree = 2)
X_transformed = Obj.fit_transform(X)
Polynomial Linear Regression Fitting
Step 4: Pre-process the independent feature using a particular degree polynomial
from sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(X_transformed, Y)
Y1= model.predict(X_transformed)
Polynomial Linear Regression Fitting
Step 5: Visualizing the fitted polynomial
plt.scatter(X,Y, color='red')
plt.plot(X,Y1, color='blue')
plt.xlabel('Level')
plt.ylabel('Salary')
plt.show()
Polynomial Linear Regression Fitting
Step 6: Analysis of the Errors
Polynomial Linear Regression Fitting
Step 6: Analysis of the Errors
from sklearn import metrics
print(metrics.mean_absolute_error(Y, Y1))
print(metrics.mean_squared_error(Y, Y1))
print(np.sqrt(metrics.mean_squared_error(Y, Y1)))
Thanks
