Lecture-5
CO4: Compare various statistical methods of study data samples
CO5: Analyze and evaluation of different sets of data using hypothesis testing.
FITTING OF EXPONENTIAL CURVE Y = abx
Let 𝑌 = 𝑎𝑏 𝑋 (1)
be an exponential curve and we have a set of n points (xi, yi), i=1, 2, ..., n. We
have to determine a and b such that equation (1) is the curve of best fit.
Taking log both sides of equation (1)
log Y = log a + log bX
log Y = log a + X log b
Let, logY = U, log a = A and log b = B
Now, equation (1) comes in the linear form as
U = A + BX (2)
which is the equation of straight line. Normal equations for equation (2) can be
obtained as
u = nA + Bx (3)
ux = A x + B x2 (4)
By solving equation (3) and equation (4), we obtain A and B and finally
a = antilog A and b = antilog B.
With these a and b, the exponential curve 𝑌 = 𝑎𝑏 𝑋 is the curve of best fit for the
given set of data.
Note: Here we are using log base 10.
Now let us solve a problem of fitting of exponential curve 𝑌 = 𝑎𝑏 𝑋 .
Example 1: Fit the exponential curve 𝑌 = 𝑎𝑏 𝑋 . from the following data.
X 2 4 6 8 10
Y 1 3 6 12 24
Solution: Let the exponential curve be 𝑌 = 𝑎𝑏 𝑋 and normal equations for
estimating a and b are
u = nA + Bx
ux = A x + B x2
where, a = antilog(A) and b = antilog (B)
x y U=log y ux x2
2 1 0.0000 0.0000 4
4 3 0.4771 1.9085 16
6 6 0.7782 4.6690 36
8 12 1.0792 8.6334 64
10 24 1.3802 13.8021 100
x=30 y=46 u=3.7147 ux=29.0130 x2=220
Substituting the values of x, y, u, ux and x2 in the above normal
equations, we get
3.7147 = 5A+30B (1)
29.0130 = 30 A+220 B (2)
Multiplying equation (1) by 6, we have
22.2882 = 30A+180B (3)
Subtracting equation (3) from equation (2), we have
29.0130 = 30A + 220B
22.2882 = 30A + 180B
--------------------------
6.7248 = 40 B
B = 0.1681
Substituting the value of B in equation (1), we get
A= − 0.26566
Thus, a = antilog A = antilog (− 0.26566) =1.8436 and
b = antilog B = antilog (0.1681) =1.4727
Thus, exponential curve of best fit is 𝑌 = 1.8436(1.4727)𝑋 .
Now let us solve a problem.
Example 2: Fit the exponential curve 𝑌 = 𝑎𝑏 𝑋 . from the following data.
X 1 2 3 4 5
Y 8 15 33 65 130
