You are on page 1of 12

STA404

Regression & Correlation Analysis

How to fit an Exponential Curve?


Procedure and step by step
calculation with example II
How to fit an Exponential Curve

Example:
Fit an exponential curve Y  a X b to data given below:

X Y

1 2.98
2 4.26
3 5.21
4 6.10
5 6.80
6 7.50
2
Transformation of given equation
Non-linear form Linear form

b
Y  aX
How …. ?

Taking log on both sides…

3
Now it is linear form …
Y   A  bX 
1st normal equation: 2nd normal equation:
(Apply  on both sides) (Multiply with X and then
Apply  on both sides)

4
Required Calculation

(Log X)*(Log Y)
X Y Log Y Log X (Log X)^2
1 2.98 2.98 0.0000 0.0000 0.0000
2 4.26 4.26 0.3010 0.0906 0.1895
3 5.21 5.21 0.4771 0.2276 0.3420
4 6.10 6.10 0.6021 0.3625 0.4728
5 6.80 6.80 0.6990 0.4886 0.5819
6 7.50 7.50 0.7782 0.6055 0.6809
Total - 4.3133 2.8574 1.7749 2.2672

5
Solving Normal Eqs…find b
Y   nA  bX     
X Y   AX   b  X      
2

6
Putting “b” … get value of “A”
Now we put the value of b = 0.515 in Eq1 and solve as

Y   nA  bX 

7
Converting back …
A  log a
 a  anti  log( A)

8
How to Use Windows Calculator
• For Anti – log • For log
– Write value – Write value
– Then Press – Then Press

9
Hence the fitted equation becomes…
b
Y  aX

Y 

10
Recap

Make suitable Get new


Given Data
Transformation equation

Solve Normal
Final Equation Converted back Eqs and Do Table
Calculations

11
- The END -

12

You might also like