1

I am new to using Eviews. I want to fit my data to an exponential curve: y = a*e^(bx). I am running the regression using these two equations:

  1. MICE_1= COEF(1)*(EXP(COEF(2)*OBSERVATION))

  2. LOG(MICE_1) = COEF(3) + COEF(4)*OBSERVATION

Since, both equations are the same, coef(2)=coef(4) and coef(1)= exp(coef(3))

However, I am getting different coefficients using the two equations. Can someone please explain why is this happening?

Mark Rotteveel
  • 100,966
  • 191
  • 140
  • 197
abhinavgnash
  • 11
  • 1
  • I think this is because the curve fitting software is minimizing something different in each case. In case 1, the error being minimized is "f1(observation) - mice_1". In case 2, the error is "f2(observation) - log(mice_1)". The second case is minimizing error based on the log of mice_1. – James Phillips Aug 04 '19 at 13:17
  • Okay, got it. If I plot my data in excel and go for an exponential trendline, the equation I get has the same coefficients I get from eqn. 2 here. That means excel is taking the log to find coefficients. Can we say that eqn1 is exponential non-linear regression method and eqn2 is exponential linear regression method? Does that make sense? – abhinavgnash Aug 05 '19 at 09:39

0 Answers0