I am using the poweRlaw package to fit some data and calculate p-values. I noticed that in some cases, where there is not so much variation in the data combined with a relatively small number of observations, I am unable to calculate a p-value once the number of simulations passes a certain threshold.
Consider the following data:
library(poweRlaw)
vec<-c(1,1,1,1,1,2,1,1,4,1,1,1,1,1,1,1,1,2,1,1,1,2,2,1,2,1,
1,2,1,1,2,1,1,1,1,1,1,2,2,2,2,1,1,1,1,2)
Can fit the power law and calculate the p-value:
pl=displ$new(vec)
est=estimate_xmin(pl)
bts=bootstrap_p(pl,no_of_sims=30,seed=42)
However, when I increase the number of simulations, to for instance 35, I get the following error:
Error in checkForRemoteErrors(val) :
one node produced an error: index 0 outside bounds
Actually I don't really understand what the error means practically. I guess an issue here is that the observed data probably doesn't follow a power law, at least not at this relatively low number of observations. But how come that this error occurs at 35 simulations but not 30?
Solution This issue was fixed in poweRlaw version 0.7.1
Run update.packages().
I think the error arises because during the bootstrap procedure, when mainly 1 are selected. This leads to numerical instability when estimating the scaling parameter.
I've created an issue at https://github.com/csgillespie/poweRlaw/issues/56