Fitting of A Modified Exponential Trend Line
Fitting of A Modified Exponential Trend Line
y3 y1- y2 2
k=
y3 - 2y2 + y1
Partial Sums
Divide the entire data into 3 equal parts and compute the partial sums for each part.
S1 = ∑nt= 1 yt
S2 = ∑2n
t = n+1 yt
S3 = ∑3n
t = 2n+1 yt
1
S3 - S2 n
b =( )
S2 - S1
b - 1 (S2 - S1)3
a= ×
b S3 - 2S2+ S1
S3 S1 - S2 2
k=
n(S3 - 2S2 + S1 )
The trend line estimates are obtained from the fitted trend line
Tt = k + abt
Calculations:
x1= 6.8429
x2= 10.0752
x3= 20.4878
b= 1.124099494
1.294422288
a=
5.387840561
k=
Partial Sums
Year Observations t Tt k
1995 6.8429 1 6.145947 3.685961
1996 7.0243 2 6.407292 3.685961
1997 7.0666 3 6.696402 3.685961
1998 7.0529 4 = S1 7.016226 3.685961
1999 6.7935 5 7.370028 3.685961
2000 7.0945 6 7.761417 3.685961
2001 7.717 7 8.194387 3.685961
2002 8.1911 8 8.673355 3.685961
2003 8.5702 9 = S2 9.203208 3.685961
2004 9.2227 10 9.789352 3.685961
2005 10.0752 11 10.43777 3.685961
2006 11.647 12 11.15507 3.685961
2007 12.6035 13 11.94857 3.685961
2008 13.724 14 12.82638 3.685961
2009 14.6939 15 13.79744 3.685961
2010 15.9344 16 14.87167 3.685961
2011 16.7902 17 16.06002 3.685961
2012 17.3739 18 = S3 17.37462 3.685961
2013 18.5253 19 18.82888 3.685961
2014 19.7821 20 20.43764 3.685961
2015 20.4878 21 22.21732 3.685961
S1 49.5917
S2 74.0337
S3 123.5876
S2 - S1 24.442
S3 - S2 49.5539
S3 - 2S2 +
S1 25.1119
b= 1.106238
a= 2.223739
k= 3.685961
Graph:
Three selected points
TIME PLOT
25
20
15
10
0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Observation T=K+abt K
Partial Sums
TIME PLOT
25
20
15
10
0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Observations Tt
Result:
The trend line has been estimated and plotted on the graph.
Tt = k + abt
Partial Sums
a = 2.2237
b = 1.1062
k = 3.6851