0

I am using the power law package in R, when we compare power law distribution with other distributions, we need to set the same Xmin. My professor asked me why we should set the same Xmin today, rather than the same C? p(x)=Cf(x), C is a normalization constant. I cannot answer this question, anyone know? Thank you so much!!!

csgillespie
  • 59,189
  • 14
  • 150
  • 185
Ian
  • 31
  • 1
  • 7

1 Answers1

2

The normalisation constant is needed to make the distribution sum/integrate to one. If we omit the constant, then we do not have a proper distribution.

The reason why the xmin values need to be idenitical is when comparing two models the data sets used to fit each model should be the same.

csgillespie
  • 59,189
  • 14
  • 150
  • 185