Correlation & Regression • CR6

Normal Equation Method to Find Regression Line

Learn how the regression line is formed from summation properties and normal equations, then practise one problem in a clear IOE-style format.

Why Normal Equation Method?

Regression line is not guessed randomly. In normal equation method, we first assume the best fitting straight line as: \[ Y=A+BX \] Then we find the values of \(A\) and \(B\) by solving two equations called normal equations.

Meaning of \(A\)

\(A\) is the intercept. It gives the predicted value of \(Y\) when \(X=0\).

Meaning of \(B\)

\(B\) is the regression coefficient. It tells how much the predicted value of \(Y\) changes when \(X\) increases by one unit.

Properties of Summation Needed Here

Property 1

\[ \sum c=nc \]

A constant \(c\) added \(n\) times gives \(nc\).

Property 2

\[ \sum c f(x)=c\sum f(x) \]

A constant outside a variable part can be taken outside the summation.

Property 3

\[ \sum [f(x)+g(x)]=\sum f(x)+\sum g(x) \]

Summation can be distributed over addition.

Derivation of Normal Equations

First Normal Equation

Assume the regression line:

\[ Y=A+BX \]

Taking summation on both sides,

\[ \sum Y=\sum(A+BX) \]

\[ \sum Y=\sum A+\sum BX \]

Using summation properties,

\[ \sum Y=nA+B\sum X \]

Therefore, the first normal equation is:

\[ \boxed{\sum Y=nA+B\sum X} \]

Second Normal Equation

Again start with:

\[ Y=A+BX \]

Multiply both sides by the independent variable \(X\):

\[ XY=AX+BX^2 \]

Now take summation on both sides:

\[ \sum XY=\sum AX+\sum BX^2 \]

Using summation properties,

\[ \sum XY=A\sum X+B\sum X^2 \]

Therefore, the second normal equation is:

\[ \boxed{\sum XY=A\sum X+B\sum X^2} \]

Short form: To find the regression line by normal equation method, solve \[ \sum Y=nA+B\sum X \] \[ \sum XY=A\sum X+B\sum X^2 \] After finding \(A\) and \(B\), write the regression line as: \[ \hat{Y}=A+BX \]
Now let us work out one problem using this method. As always, try it first. If you are not able to complete it, open the detailed solution or watch the video solution. Do not forget to support SJMathTube.

Problem: Speed Reading Programme

Normal Equation Method • Regression Line • Prediction

The following data gives the number of weeks of practice in a speed reading programme and the corresponding reading speed score. Fit the regression line of reading speed score on weeks of practice by normal equation method. Also estimate the reading speed score after \(6\) weeks of practice.

Weeks of Practice \(X\) 1 2 3 4 5
Reading Speed Score \(Y\) 52 55 61 66 70
✓
Enter your answer first.
Correct answer: \(46.70\)
✓
Enter your answer first.
Correct answer: \(4.70\)
✓
Enter your answer first.
Correct answer: \(74.90\)
✓
Select your answer first.
Correct answer: Predicted score increases by about \(4.70\) for each additional week of practice.

Try once before opening the solution

First prepare the table for \(\sum X\), \(\sum Y\), \(\sum X^2\), and \(\sum XY\). Then form the two normal equations and solve for \(A\) and \(B\).

Detailed Step-by-Step Solution

Let weeks of practice be \(X\) and reading speed score be \(Y\).

\(X\) \(Y\) \(X^2\) \(XY\)
1 52 1 52
2 55 4 110
3 61 9 183
4 66 16 264
5 70 25 350
Total \(\sum X=15\) \(\sum Y=304\) \(\sum X^2=55\), \(\sum XY=959\)

Assume the regression line:

\[ Y=A+BX \]

The normal equations are:

\[ \sum Y=nA+B\sum X \]

\[ \sum XY=A\sum X+B\sum X^2 \]

Substitute the values:

\[ 304=5A+15B \]

\[ 959=15A+55B \]

Solving the two equations,

\[ A=46.70,\qquad B=4.70 \]

Therefore, the regression equation is:

\[ \hat{Y}=46.70+4.70X \]

For \(6\) weeks of practice, put \(X=6\):

\[ \hat{Y}=46.70+4.70(6) \]

\[ \hat{Y}=74.90 \]

The estimated reading speed score after \(6\) weeks of practice is \(74.90\).

Meaning of the regression coefficient:

The value \(B=4.70\) means that for each additional week of practice, the predicted reading speed score increases by about \(4.70\).

Final Answer and Interpretation

Regression equation: \(\hat{Y}=46.70+4.70X\) Predicted reading speed score for \(6\) weeks of practice \(=74.90\) The regression coefficient \(4.70\) means predicted reading speed score increases by about \(4.70\) for each additional week of practice.

Video Solution

In this video, we solve the regression line using the normal equation method.
Formulae

Available Formula Sheets

⌂ Home
Latest
Latest Content Auto
Home