Polynomial Regression

x1.02.03.04.0
y6.011.018.027.0
regression table data
Advertisements

Solution:

Let Y = a1 + a2x + a3x2    ( 2nd order polynomial ).

Here, m = 3  ( because to fit a curve we need at least 3 points ).

Since the order of the polynomial is 2, therefore we will have 3 simultaneous equations as below.

Polynomial Regression Formula

To learn more, see what is Polynomial Regression

polynomial regression formula
Advertisements

Evualuate Formula

Now we have to evaluate the values which are required according to the above equations.

Let us find the value of x2, x3, x4, yx, and yx2.

xyx2x3x4yxyx2
1.06.011166
2.011.049162244
3.018.09168154162
4.027.01636256108432
xi=10yi=62x2=364x3=524x4=354yx=190∑yx2=644
Regression table

Put the values in the above 3 equations as below.

4a1 + 10a2 + 30a3 = 62   …. ( 1 )

10a1 + 30a2 + 100a3 = 190  …. ( 2 )

30a1 + 100a2 + 354a3 = 644  … ( 3 )

In order to solve the above 3 simultaneous equations, we will write the above equations in the form of matrices as below.

Fit a Second Order Polynomial to the  following given data.
Fit a Second Order Polynomial
Fit a Second Order Polynomial to the  following given data. Curve fitting Polynomial Regression using gauss elimination method

Now by using back substitution we can find the values of a1, a2, and a3.
Here, 4a3 = 4 , a3 = 1
And, a2 + 5a3 = 7 , a2 = 2
Also, a1 + 2.5a2 + 7.5a3 = 15.5 , a1 = 3
Therefore, a1 = 3, a2 = 2, a3 = 1.

The Quadratic equation becomes:
Y = 3 + 2x + x2, or

Y = x2 + 2x + 3

Suggestion:

  1. Newtons Interpolation
  2. Lagrange’s Interpolation
  3. Bisection method
  4. Regula falsi method
  5. Linear Regression

Advertisements