1

I watched this video on Fisher Information and it is mentioned that in Taylor series expansion of the likelihood function the second derivative is parabola which is not a good approximation and a better approximation is second derivative of log of likelihood function which is gaussian.

Question: Why gaussian (second derivative of log of likelihood) is a better approximation then parabola (second derivative of likelihood)?

GENIVI-LEARNER
  • 720
  • 4
  • 13
  • See https://stats.stackexchange.com/questions/289190/theoretical-motivation-for-using-log-likelihood-vs-likelihood – Henry Apr 06 '20 at 12:09

0 Answers0